Episode 69

full
Published on:

6th Feb 2025

69. Building an AI First Organisation

In this episode we discuss: Making AI happen in 2025 for your organization. We are joined by Charlie Cowan, Author of "How To Sell Tech" and “The Revenue Operations Playbook”.

Love The Operations Room? Please support us by rating and reviewing it here.

We chat about the following with Charlie Cowan: 

  • How can professionals stay updated in an era where AI is rapidly disrupting industries like sales and marketing?
  • With AI advancing faster than its everyday adoption, how can businesses bridge the gap between innovation and practical implementation?
  • How can non-developers leverage AI tools to accelerate product development while overcoming emotional and technical challenges?
  • In a world where restrictive AI policies hinder adoption, how can organisations balance data privacy concerns with fostering innovation?
  • Rather than replacing jobs, how can AI be used to supercharge teams, enhance leadership effectiveness, and drive productivity?

References 

  • https://www.linkedin.com/in/charliecowan/
  • charliecowan.ai
  • https://v0.dev/
  • https://replit.com/ai
  • https://lovable.dev/
  • Bolt.new
  • https://www.cursor.com/
  • https://codeium.com/windsurf
  • https://notebooklm.google.com/

Biography 

Charlie Cowan helps organisations accelerate AI adoption, guiding CxOs in embedding AI-driven processes to unlock new opportunities. As the founder of Kowalah, an AI-powered buying platform, he built the business from scratch—without prior coding experience—using AI tools. Now, he shares his journey to inspire others to embrace AI innovation.

An author of four books on sales, revenue operations, and go-to-market strategy, Charlie provides practical insights for scaling startups and sales teams. Passionate about AI and business transformation, he continues to drive conversations on the future of AI adoption and leadership.

To learn more about Beth and Brandon or to find out about sponsorship opportunities click here. 

Summary

20:13 Introduction to Charlie Cowan and His Journey

23:14 The Impact of AI on Business and Personal Reinvention

25:35 Building Koala: The Journey of a Non-Developer

28:00 Navigating Challenges in AI Development

30:48 Balancing Consulting and Product Development

31:27 Leveraging LinkedIn for AI Insights

32:40 The AI Bubble and Company Policies

34:55 Embracing AI: Opportunities and Risks

41:28 Transforming Organisations with AI

44:02 Innovative Tools for Information Management

46:38 Practical AI Applications in Leadership

49:23 Final Thoughts on AI and Automation



This podcast uses the following third-party services for analysis:

Podcorn - https://podcorn.com/privacy
Transcript
Speaker:

Hello everyone, and welcome to

Speaker:

another episode of The Operations

Speaker:

Room, a podcast for CEOs.

Speaker:

I am Brandon Bensinger, joined by my

Speaker:

amazing costar Bethany Errors.

Speaker:

How are things going, Bethany?

Speaker:

I don't know.

Speaker:

So one thing is I don't

Speaker:

think Instagram is good for me.

Speaker:

I suspect Instagram isn't good for

Speaker:

anybody, but yet despite that,

Speaker:

I still am on it

Speaker:

and I'm still on it probably too

Speaker:

much.

Speaker:

And it really feeds my health

Speaker:

anxiety.

Speaker:

Health news and all the ways

Speaker:

that we are killing ourselves and

Speaker:

all of the rules that I'm not

Speaker:

following and all

Speaker:

of the reasons why when I die, it's

Speaker:

going to be my fault.

Speaker:

And even though I'm aware of it and

Speaker:

I have blocked as many health

Speaker:

news things as I

Speaker:

can, any time it

Speaker:

comes up, I can't help but watch.

Speaker:

And then the algorithms like, what

Speaker:

she cares about is health news.

Speaker:

And then it just comes back in.

Speaker:

So the most recent

Speaker:

one was actually not specifically

Speaker:

a health news

Speaker:

influencer account.

Speaker:

It was a data account

Speaker:

and it was about how women

Speaker:

are I don't know if it's dying of

Speaker:

cancer or just getting cancer

Speaker:

at a tremendously higher rate

Speaker:

than men.

Speaker:

And then you had comments from

Speaker:

random people trying to decide

Speaker:

why it is that women are getting

Speaker:

cancer at much higher rates than

Speaker:

men. And those comments have seeped

Speaker:

into my brain and now I'm

Speaker:

freaking out.

Speaker:

So it was that

Speaker:

women wear a lot

Speaker:

more cosmetics than men

Speaker:

and they're not properly tested and

Speaker:

a lot of them are made with

Speaker:

petroleum products.

Speaker:

And so we're just like smothering

Speaker:

oil all over our bodies.

Speaker:

And that causes issues.

Speaker:

Men tend to wear more

Speaker:

cotton than women, and women wear

Speaker:

more special fabrics, and all

Speaker:

of the special fabrics are made

Speaker:

out of plastic.

Speaker:

And as they degrade, you get

Speaker:

microplastics in your body,

Speaker:

sports bras directly

Speaker:

linked to breast cancer.

Speaker:

I mean, all of this is not real.

Speaker:

And these are people in the comment

Speaker:

section postulating

Speaker:

as to why women have higher rates of

Speaker:

cancer.

Speaker:

But my ability

Speaker:

to cheat death and do

Speaker:

all of the right things,

Speaker:

follow all the rules

Speaker:

and never die.

Speaker:

That's a lot of pressure on

Speaker:

somebody's shoulders.

Speaker:

At some point you will die,

Speaker:

Bethenny, I suspect.

Speaker:

No, no. If I follow all the rules,

Speaker:

I'll be fine forever.

Speaker:

So sometimes Instagram's not bad.

Speaker:

I was served a video

Speaker:

earlier this week That was Jane

Speaker:

Fonda talking about old age.

Speaker:

And the opening of the video was the

Speaker:

best because she said,

Speaker:

I am Jane Fonda.

Speaker:

You may know me from being Jane

Speaker:

Fonda for the last 86 years.

Speaker:

Exactly. And growing up, it's

Speaker:

literally impossible somehow to not

Speaker:

know who she is.

Speaker:

In the video, she it's like really.

Speaker:

Well, production is video.

Speaker:

And she's working with a scientist

Speaker:

in the scientist is talking about

Speaker:

the facts. And then Jane Fonda is

Speaker:

adding in her bits.

Speaker:

And one of the things she said like

Speaker:

trying to banish senior moments

Speaker:

as a thing.

Speaker:

And it's just like what we're

Speaker:

telling ourselves is that we're

Speaker:

losing our memory. Like when we

Speaker:

can't find our car keys when we're

Speaker:

17.

Speaker:

We're not saying, it's a senior

Speaker:

moment, we just can't find our car

Speaker:

keys. And to embrace

Speaker:

aging and not be so afraid of it

Speaker:

and also say that we all focus

Speaker:

on ending up in nursing homes

Speaker:

and freaking out about nursing

Speaker:

homes.

Speaker:

But actually only 3%

Speaker:

of the population will end up in

Speaker:

them for that.

Speaker:

True? 3%.

Speaker:

I mean, that's true according to

Speaker:

Instagram and Jane Fonda.

Speaker:

I have no idea if that's true.

Speaker:

We could go ahead and verify that.

Speaker:

Jane Fonda said it has got to be

Speaker:

true.

Speaker:

Jane Fonda and the scientist

Speaker:

said it. So, you know,

Speaker:

I choose to believe that that is

Speaker:

true, that 3% of us are going to end

Speaker:

up in nursing homes and the rest of

Speaker:

us are going to have some

Speaker:

combination of home care or just

Speaker:

dying doing okay.

Speaker:

So anyhow, that's the inner workings

Speaker:

of my brain. So how am I doing,

Speaker:

Brandon? I don't know.

Speaker:

All right. So we have got an amazing

Speaker:

topic for today, which is making AI

Speaker:

happen in 2025 for your

Speaker:

organization.

Speaker:

We have an amazing guest for this,

Speaker:

which is Charlie Cullen.

Speaker:

He is an AI strategist and

Speaker:

has singlehandedly developed a

Speaker:

product called Koala using

Speaker:

AI tools behind him in terms

Speaker:

of development tools to make the app

Speaker:

happen. Being a non techie, which is

Speaker:

phenomenal and quite an amazing

Speaker:

story.

Speaker:

So before we get to Charlie, the

Speaker:

first question I wanted to ask you

Speaker:

bit of a broad question, which is

Speaker:

what is the opportunity right now

Speaker:

for scale ups to go horror

Speaker:

on being AI first, as it were.

Speaker:

I made a face when you asked that

Speaker:

question because I was wondering if

Speaker:

you meant I first in the product

Speaker:

or I first within the company.

Speaker:

I mean, I think there's the bigger

Speaker:

existential question is, is

Speaker:

your product shit and do you need to

Speaker:

tear it up and just started again as

Speaker:

an AI product.

Speaker:

And if you're more on the startup

Speaker:

and scale up side, I would highly

Speaker:

encourage you to do that.

Speaker:

I'm just pushing that one out.

Speaker:

There is like

Speaker:

a little moment of of

Speaker:

compelling thinking or controversial

Speaker:

thoughts, although it's interesting

Speaker:

because everything is so new.

Speaker:

There's a lot of scope for improved

Speaker:

products. Like for me, the really

Speaker:

big gap is

Speaker:

a next generation Zapier

Speaker:

like something that is really easy

Speaker:

for non-technical users to you

Speaker:

to use to automate

Speaker:

the soul destroying work in

Speaker:

their lives.

Speaker:

And there are loads and loads of

Speaker:

companies trying to figure out

Speaker:

a tick, but they're still like

Speaker:

technical for technical people

Speaker:

rather than

Speaker:

no code.

Speaker:

Easy automation for

Speaker:

non techies.

Speaker:

So I'm very excited about that

Speaker:

space. And once that.

Speaker:

Unlocks.

Speaker:

Then everything in your business

Speaker:

should be your first.

Speaker:

It's a bit of a question of agents

Speaker:

which are going to be verticals

Speaker:

right now for very specific

Speaker:

functions as opposed to agents

Speaker:

working together in tandem to

Speaker:

actually have more of a net effect

Speaker:

for an organization.

Speaker:

And I think that's definitely out

Speaker:

there in the future. I think we're

Speaker:

going to get first is more siloed

Speaker:

agents for Realm A

Speaker:

versus Run B?

Speaker:

Yeah, but like a lot of those agents

Speaker:

are just siloed.

Speaker:

Vertical agents are not interesting

Speaker:

for me. My Nirvana.

Speaker:

And I think at some point this year

Speaker:

will be I

Speaker:

have all of the CVS to sort

Speaker:

through.

Speaker:

I don't want to sort through them.

Speaker:

I tell the agent, these

Speaker:

are the topics that I'm looking for

Speaker:

on the CVS.

Speaker:

This is what good looks like.

Speaker:

Go and read all of these and

Speaker:

bring out the ones

Speaker:

that I'm looking for rather than a

Speaker:

rules based.

Speaker:

And then you're like, No, no, you

Speaker:

didn't quite get the CVS, right?

Speaker:

Really? I want to have blah, blah,

Speaker:

blah, blah, blah, blah. And, you

Speaker:

know, like actually doing

Speaker:

the boring work for you, but

Speaker:

you're prompting them in English.

Speaker:

But I would consider that to be a

Speaker:

vertical sized agent for

Speaker:

the people recruitment function as

Speaker:

an example.

Speaker:

I think that's definitely in sight

Speaker:

right now. I think what I was

Speaker:

thinking more in my head is more

Speaker:

around broader activities

Speaker:

where, you know, it's not just

Speaker:

simply a recruiter and CV thing that

Speaker:

they're doing, but in fact it's

Speaker:

maybe the entire people function.

Speaker:

Multiple agents need to work

Speaker:

together and cross-pollinate to

Speaker:

come up with better outcomes, I

Speaker:

suppose.

Speaker:

I think you're right in terms of

Speaker:

exists now. But the problem is, is

Speaker:

every single one of those things

Speaker:

exists as a $10 subscription

Speaker:

and I want to pay $130

Speaker:

subscription and do it all

Speaker:

where, you know, there's like one

Speaker:

tool that the people can use

Speaker:

to do their stuff and

Speaker:

accounts payable can use to do

Speaker:

accounts payable and sales

Speaker:

can use to automate their stuff.

Speaker:

And you have like a new

Speaker:

automation layer that the

Speaker:

entire business can

Speaker:

use.

Speaker:

I think from an operational

Speaker:

standpoint, for a company, it seems

Speaker:

like the low hanging fruit

Speaker:

that is out there is simply your

Speaker:

processes and policies.

Speaker:

Anything that's been documented in

Speaker:

that form, which is either

Speaker:

internally creating a subset of GB

Speaker:

TS that pass off different pieces

Speaker:

of policy sections or process

Speaker:

sections or functions or whatever.

Speaker:

Or what I've also seen is

Speaker:

using GB to your court or what have

Speaker:

you. But what I've also seen is

Speaker:

actual agents that

Speaker:

are set up to ingest your

Speaker:

policies or ingest your processes

Speaker:

and do something more advanced than

Speaker:

what you get with an internal chat

Speaker:

and GPT that you create yourself

Speaker:

basically.

Speaker:

So one example of this was

Speaker:

the internal knowledge base curator,

Speaker:

GPT.

Speaker:

So that product and it is a product,

Speaker:

you basically take your

Speaker:

documentation with your company,

Speaker:

ingest it into the CBT.

Speaker:

It's all set up to sort, categorize

Speaker:

and provide a more advanced

Speaker:

experience for the end user in terms

Speaker:

of being able to search for things

Speaker:

and so on with a taxonomy that's

Speaker:

created and so on.

Speaker:

And that seems very interesting to

Speaker:

me.

Speaker:

Yeah, I just don't want to have to

Speaker:

pay 10 pounds for every single one

Speaker:

of them or 20 pounds for every

Speaker:

single one of them.

Speaker:

I'm looking forward to the

Speaker:

consolidation and it just at

Speaker:

what level does that consolidation

Speaker:

happen?

Speaker:

So I was reading Charlie's

Speaker:

LinkedIn posts and he created one

Speaker:

post around hiring

Speaker:

for an AI first organization,

Speaker:

and his little description that he

Speaker:

put in there was to create a task

Speaker:

for a candidate where

Speaker:

they need to write a prompt or

Speaker:

create a. Claude Church project

Speaker:

that can accelerate a key task

Speaker:

in the role that they're being hired

Speaker:

for and to demonstrate

Speaker:

an example of how they would use it.

Speaker:

And thus that's really interesting

Speaker:

to me. Or maybe think at the very

Speaker:

least in the sense of like, okay, if

Speaker:

we want to hire people

Speaker:

that are on a I using

Speaker:

a, I have some level of competence

Speaker:

and skill and interest in A.I..

Speaker:

It seems like at the very least

Speaker:

asking an interview question in some

Speaker:

form, maybe not a task, but some

Speaker:

kind of questioning around this

Speaker:

seems to make tremendous sense.

Speaker:

I'm just curious what you think.

Speaker:

I agree. And I actually think being

Speaker:

a task would be totally legitimate.

Speaker:

And it shows interest.

Speaker:

It shows aptitude and it

Speaker:

shows curiosity.

Speaker:

I think there's still a level of

Speaker:

resistance to

Speaker:

a AI because it's seen

Speaker:

as cheating.

Speaker:

And so it's actually showing people

Speaker:

who have a AI first

Speaker:

mindset or forward mindset

Speaker:

where using A.I.

Speaker:

isn't cheating, it's the new normal

Speaker:

and it's the smart thing to do.

Speaker:

And so both you're testing

Speaker:

for people in who

Speaker:

are coming into your organization,

Speaker:

who are afterward, but

Speaker:

you're also massively demonstrating

Speaker:

expectations that in your

Speaker:

world, using

Speaker:

AI is smart, not cheating.

Speaker:

So when you think about rolling

Speaker:

out A.I.

Speaker:

within the organization right now in

Speaker:

2025 for a company, we

Speaker:

have our standard tools, if you want

Speaker:

to call it that, which is clod

Speaker:

chatbot, Gemini.

Speaker:

What makes sense here?

Speaker:

Would you encourage companies right

Speaker:

now to say to themselves, okay,

Speaker:

look, let's create a policy around

Speaker:

this. Let's license

Speaker:

some of these tools for different

Speaker:

functions within the company or

Speaker:

within the entire organization and

Speaker:

launch them into the company.

Speaker:

Upskill people in terms of how to

Speaker:

write prompts or create projects,

Speaker:

create some level of support

Speaker:

structure around it where people are

Speaker:

being supported to use it

Speaker:

effectively, and some kind of reward

Speaker:

mechanism to reward those who are

Speaker:

actually integrating it back into

Speaker:

their jobs.

Speaker:

Is that something that we should do

Speaker:

now?

Speaker:

All the technology is changing so

Speaker:

rapidly that I would

Speaker:

not commit to a single one.

Speaker:

I would go even if it's a bit more

Speaker:

expensive for rolling monthly

Speaker:

contracts and continue to

Speaker:

experiment.

Speaker:

It's too early

Speaker:

to sign up for an annual

Speaker:

commitment to anything right now.

Speaker:

I don't even know if you can, but if

Speaker:

you can, I would not suggest annual

Speaker:

commitments.

Speaker:

It's not worth it.

Speaker:

Things are moving fast.

Speaker:

Policies, Definitely.

Speaker:

Because it makes you think through

Speaker:

all of the gotchas and it makes it

Speaker:

really clear for the organization

Speaker:

just to push

Speaker:

Charlie's LinkedIn a little bit

Speaker:

more. He happened to have a LinkedIn

Speaker:

post this morning that was a seven

Speaker:

page, a AI policy

Speaker:

template to use that looked pretty

Speaker:

decent for what a lot of

Speaker:

the main thoughts if you want, if

Speaker:

you haven't put one in place yet or

Speaker:

you want to make sure that you're

Speaker:

covering the key areas, I suggest

Speaker:

having a look at it.

Speaker:

And then the reward mechanism or

Speaker:

for us at peak,

Speaker:

we have a few different ways that

Speaker:

we are focusing on the team

Speaker:

using and getting more experimental

Speaker:

and interested in generally because

Speaker:

we have pockets that are super

Speaker:

interested in pockets that aren't.

Speaker:

So in our weekly

Speaker:

we have like a company all hands

Speaker:

that's weekly.

Speaker:

There is a five minute

Speaker:

segment every single week where

Speaker:

somebody does a show and tell on

Speaker:

something they've done.

Speaker:

The second thing

Speaker:

that we've done is we

Speaker:

have a Slack channel that's a

Speaker:

Jenn-air channel. We actually used

Speaker:

to have two and have just merged it

Speaker:

into one because it was getting a

Speaker:

bit messy as to which is which.

Speaker:

And so all news, all

Speaker:

thoughts, all new technology,

Speaker:

all uses goes

Speaker:

into that channel and

Speaker:

it's a self-selecting channel.

Speaker:

But I think most of the businesses

Speaker:

in it.

Speaker:

And then the third thing that we are

Speaker:

rolling out this year is

Speaker:

for our technical teams,

Speaker:

a Jenn-air training course

Speaker:

we're running internally.

Speaker:

We've not found anything externally

Speaker:

that matches our needs and

Speaker:

that is an enablement course.

Speaker:

It'll be running for the first half

Speaker:

of the year and then we'll see what

Speaker:

the second half brings us.

Speaker:

So when it comes to budgeting

Speaker:

and providing different

Speaker:

functions that have different

Speaker:

aspirations and different needs

Speaker:

around some of these generic tools,

Speaker:

what are you doing from a budgeting

Speaker:

standpoint? You're you're creating

Speaker:

like a budget for each of the

Speaker:

functions or how does that work?

Speaker:

So we're doing a bit of of that

Speaker:

for experimentation and like what

Speaker:

are the right tools, what are the

Speaker:

best tools?

Speaker:

And then also for all

Speaker:

of our systems of records

Speaker:

or standard SAS tools

Speaker:

as they come up for renewal.

Speaker:

We are investigating

Speaker:

the market and looking for

Speaker:

eye first alternatives

Speaker:

and seeing whether or not they're

Speaker:

too early stage to move to.

Speaker:

But our general

Speaker:

assumption of hypothesis is

Speaker:

all of the SAS businesses

Speaker:

that have a bit of AI stuck

Speaker:

on the side are not going to

Speaker:

be as good as AI first businesses.

Speaker:

And so we're constantly

Speaker:

searching in the market for what

Speaker:

is the new system of record.

Speaker:

Like in this new world.

Speaker:

And for my vision, I think we have

Speaker:

one system of record with lots of

Speaker:

different elements to it.

Speaker:

Because what's a system?

Speaker:

A record, but a massive database.

Speaker:

And so and there's always

Speaker:

been that the reason why the systems

Speaker:

record are different is because of

Speaker:

the interface and the tooling

Speaker:

required.

Speaker:

But if the interface becomes

Speaker:

a chat box

Speaker:

for pretty much everybody,

Speaker:

do we need different systems of

Speaker:

record long term?

Speaker:

Short term we do.

Speaker:

Because I don't think anybody's

Speaker:

reached that long term goal.

Speaker:

But that's part of why we're not

Speaker:

committing to

Speaker:

long term contracts right now while

Speaker:

we watch the space.

Speaker:

And so for our existing systems of

Speaker:

record, when they're coming up for

Speaker:

renewal, we're not doing three year

Speaker:

renewals for anything.

Speaker:

We're doing one year renewals so

Speaker:

that we have this choice, even if

Speaker:

nothing's fit for purpose.

Speaker:

Yet our our guest is something will

Speaker:

be fit for purpose in a year.

Speaker:

So when it comes to the policy

Speaker:

that you now have and you also

Speaker:

mentioned Charlie on his LinkedIn

Speaker:

profile, putting together a

Speaker:

template, a policy.

Speaker:

He talked about a couple of

Speaker:

interesting things. One was in the

Speaker:

policy itself to provide

Speaker:

a list of encouraged uses for

Speaker:

each function that are endorsed by

Speaker:

the leaders.

Speaker:

And I think this is quite important

Speaker:

and useful in the sense that the

Speaker:

policy itself is not simply there

Speaker:

to tell people what to not do, but

Speaker:

in fact to do the opposite in this

Speaker:

case, because what we want with AI

Speaker:

is to truly embed in

Speaker:

people's thoughts that we want the

Speaker:

company to win.

Speaker:

And the way the company is going to

Speaker:

win of the future is using A.I.

Speaker:

tools. And we need you to try them.

Speaker:

We need you to experiment.

Speaker:

And by providing a list of

Speaker:

encouraged use cases and

Speaker:

it's not to say that it's going to be

Speaker:

exhaustive, but you need to put

Speaker:

something out there where people

Speaker:

like, okay, within my function of

Speaker:

sales, here's like the ten things

Speaker:

that the cells like right now has

Speaker:

tacitly endorsed, basically.

Speaker:

So maybe I should start doing

Speaker:

something about it, I guess.

Speaker:

What do you think of that first one?

Speaker:

So we have that, but not

Speaker:

in our policy.

Speaker:

We have it in our strategy document.

Speaker:

And so what we've done is

Speaker:

for each department, we

Speaker:

have like low hanging fruit,

Speaker:

you know, what are like the immediate

Speaker:

things that we can be doing.

Speaker:

What are areas where we want to

Speaker:

do more, investigate and experiment

Speaker:

with, and what is our long term

Speaker:

vision?

Speaker:

So what is our ideal in each of

Speaker:

those areas?

Speaker:

What we have also done that I forgot

Speaker:

to mention is we have an A.I.

Speaker:

evangelist in every team.

Speaker:

So that's the person

Speaker:

who self-selected

Speaker:

could be very different in each

Speaker:

team, but it's the person who just

Speaker:

like naturally is reading

Speaker:

everything, naturally experimenting.

Speaker:

They tend to be the person who's

Speaker:

in the all hands showing

Speaker:

information, but they're like the go

Speaker:

to person of Ah,

Speaker:

I've just realized that stars can

Speaker:

do whatever.

Speaker:

Or in the

Speaker:

second one that he suggested should

Speaker:

be in the policy is deciding

Speaker:

on a list of company approved A.I.

Speaker:

tools in a particular a streamlined

Speaker:

process to get tools and versions

Speaker:

approved, i.e.

Speaker:

somebody see something, they make a

Speaker:

request, they get an approval within

Speaker:

days as an SLA.

Speaker:

And the default is to say yes,

Speaker:

pretty much.

Speaker:

What do you make of that?

Speaker:

So we already have that in place,

Speaker:

but not specifically for genitals,

Speaker:

just in general for buying

Speaker:

for our security reasons.

Speaker:

So it has to go through a security

Speaker:

audit, which means that it's not

Speaker:

within a couple hours.

Speaker:

It just depends on how much

Speaker:

information we can get about these

Speaker:

companies and how they're treating

Speaker:

our data and whether or not they

Speaker:

have ISO 27,001,

Speaker:

whether or not they're GDPR

Speaker:

compliant, etc.

Speaker:

And sometimes that can take weeks.

Speaker:

But I think it's worth

Speaker:

it because there's so many new tiny

Speaker:

companies that

Speaker:

the chat you or

Speaker:

the Open Eyes and Gemini

Speaker:

like you can find out

Speaker:

all of their security really quickly

Speaker:

and make decisions and understand

Speaker:

whether or not your data is training

Speaker:

the models, etc..

Speaker:

Tiny company with 15

Speaker:

people, we end up contacting

Speaker:

them, getting their certifications,

Speaker:

understanding what their policies

Speaker:

all are. If they don't have certifications

Speaker:

yet, and talking through whether or

Speaker:

not we feel comfortable working with

Speaker:

them. Quite often they choose small

Speaker:

and haven't not mature enough and

Speaker:

don't have it yet, or maybe haven't

Speaker:

even been around enough for an

Speaker:

audit.

Speaker:

But we will have conversations with

Speaker:

them and ask them how they're

Speaker:

thinking about things and if

Speaker:

they seem to understand what they're

Speaker:

talking about and be quite forward

Speaker:

thinking and have policies in

Speaker:

place and a roadmap of where they're

Speaker:

going to get to, depending on the

Speaker:

sensitivity of the data that might

Speaker:

be going there will still say

Speaker:

yes.

Speaker:

But if we talk to somebody and it's

Speaker:

just like, okay, these people are

Speaker:

very immature, they have no idea.

Speaker:

We cannot trust them with any level

Speaker:

of data, we'll say no.

Speaker:

The other thing that occurs to me

Speaker:

for companies right now,

Speaker:

irrespective of any policy, are

Speaker:

actually rolling on AI at all

Speaker:

is just this obvious thing where

Speaker:

right now today, if you don't have

Speaker:

any AI policy, you've got a problem

Speaker:

a little bit because people in your

Speaker:

company are using personal accounts

Speaker:

to use chatbot to write their emails

Speaker:

or whatever that is happening.

Speaker:

So having a very simple statement to

Speaker:

the company, irrespective of

Speaker:

anything else, just saying, look, if

Speaker:

you're using a personal.

Speaker:

Howard, please, please, please

Speaker:

go into your settings and toggle

Speaker:

the data learning off essentially,

Speaker:

because if you don't, anything that

Speaker:

you do will be sucked into the

Speaker:

vortex of the the

Speaker:

large language models and we

Speaker:

cannot have that.

Speaker:

And also then write an AI policy.

Speaker:

The last bit I wanted to bring up

Speaker:

was best practices.

Speaker:

And you had a couple of nuggets here

Speaker:

that I thought were useful.

Speaker:

One was when it comes to

Speaker:

clod projects and

Speaker:

church EPD projects and projects

Speaker:

are essentially a framework to

Speaker:

allow you to have a consistent set

Speaker:

of context for whatever

Speaker:

it is that you're asking of the GPU

Speaker:

in this case to respond to around

Speaker:

the company or around your products

Speaker:

or around your policies or whatever,

Speaker:

whereby you don't have to enter

Speaker:

a prompt every single time to

Speaker:

put in that context.

Speaker:

And his best

Speaker:

practice was to ensure that

Speaker:

those projects and those statements

Speaker:

around the company were there

Speaker:

to allow any user within the company

Speaker:

to take that project, take that

Speaker:

context and apply it to their

Speaker:

their account very specifically

Speaker:

to ensure that all the context was

Speaker:

all pre set up.

Speaker:

When you think about one.

Speaker:

That is beyond what I've experienced

Speaker:

so far. So I just be like, if

Speaker:

Charlie says it's a good idea, it's

Speaker:

a good idea.

Speaker:

So what do we park it here and get

Speaker:

on to our conversation with Mr.

Speaker:

Charlie Cullen?

Speaker:

Why don't you talk a little bit

Speaker:

about Koala?

Speaker:

Not so much for a product that

Speaker:

people should go and buy, but the

Speaker:

journey of creating it.

Speaker:

So in August 2024,

Speaker:

this is when I was in pink.

Speaker:

My goodness. I'm going to set up a

Speaker:

i boutique consultancy.

Speaker:

And then suddenly, my

Speaker:

goodness. Like, who am I to go and

Speaker:

tell people how

Speaker:

they should be implementing and

Speaker:

integrating these technologies if

Speaker:

I've not done it myself?

Speaker:

That's just snake oil.

Speaker:

This was August.

Speaker:

I was on a ferry to a Greek

Speaker:

island with my family,

Speaker:

and the kids were sleeping

Speaker:

and my wife and I were reading

Speaker:

and I was reading a book called The

Speaker:

Jolt Effect.

Speaker:

And they did some research where

Speaker:

they looked at why enterprise

Speaker:

sales end up with the status

Speaker:

quo.

Speaker:

And the prevailing thought was that

Speaker:

people stayed with the status quo

Speaker:

because as a sales person,

Speaker:

you had not been able to sell

Speaker:

the difference from what they've

Speaker:

got today.

Speaker:

So I came back from this holiday and

Speaker:

I was thinking, okay, well, I'm

Speaker:

going to build an AI type solution

Speaker:

for this and it will solve my

Speaker:

problem of I've never built

Speaker:

anything with these tools and then

Speaker:

I'll be able to advise

Speaker:

people.

Speaker:

And even in September

Speaker:

of 2024, I

Speaker:

was WhatsApp paying friends.

Speaker:

Do you know anyone that is a

Speaker:

developer that could know how to

Speaker:

write this thing? Because I'm a

Speaker:

non-technical founder.

Speaker:

Does anyone know anyone that could

Speaker:

help me either as an agency

Speaker:

or whatever? But just over the

Speaker:

course of the summer, some amazing

Speaker:

new tools had been coming out

Speaker:

new versions of Code,

Speaker:

which is one of my favorite albums,

Speaker:

which is like a

Speaker:

a new tool called Cosa came

Speaker:

out, which is an AI development

Speaker:

environment.

Speaker:

A new tool called V Zero

Speaker:

came out, which is a UI

Speaker:

development tool.

Speaker:

And I just thought one weekend, you

Speaker:

know, I'll just see how

Speaker:

far I can get as a non

Speaker:

developer.

Speaker:

And so I started off with

Speaker:

Claude.

Speaker:

So if you use EPG, then you

Speaker:

know how to use Claude.

Speaker:

One of my favorite features of Tord

Speaker:

is that you can set up a project

Speaker:

and a project you can think of as a

Speaker:

wrapper around a specific subject

Speaker:

or topic.

Speaker:

So I set up a project

Speaker:

called Koala, which is the

Speaker:

name of the application that I

Speaker:

built, and I gave it some custom

Speaker:

instructions. You are my technical

Speaker:

co-founder and sure

Speaker:

I am. Got a clue what I'm doing.

Speaker:

I've got an idea for an app, but I

Speaker:

want you to be my co-founder

Speaker:

and to help me to architect it, to

Speaker:

help me structure the project,

Speaker:

to help me to understand what

Speaker:

tech stack I should use.

Speaker:

And I'm going to come back to you

Speaker:

with questions and you should

Speaker:

push back on me if what I'm

Speaker:

asking for is a stupid

Speaker:

idea. And our goal here

Speaker:

is just to get to an MVP

Speaker:

where we can get some paying

Speaker:

customers.

Speaker:

So there's more that goes into it

Speaker:

than that. But that was basically

Speaker:

the custom instruction.

Speaker:

And then as I started chatting

Speaker:

with Claude, I would build out

Speaker:

Google Docs. That might be the

Speaker:

project overview, it might be the

Speaker:

file structure, it might be the

Speaker:

licensing model, it might

Speaker:

be some of the go to market.

Speaker:

And all of this becomes

Speaker:

the context that is in this flawed

Speaker:

project.

Speaker:

And my goodness, I chat

Speaker:

to this thing all day long

Speaker:

at the weekends because this is a

Speaker:

weekend project from eight

Speaker:

in the morning I'm chatting to

Speaker:

Claude, say, Hey, Claude, you know

Speaker:

what you think about this. What do

Speaker:

you think about that?

Speaker:

And one of the great things

Speaker:

about having Claude as a technical

Speaker:

co-founder is you can ask

Speaker:

stupid questions.

Speaker:

You can ask the question that you

Speaker:

asked Claude yesterday, and, you

Speaker:

know, he doesn't complain.

Speaker:

And because you give him the

Speaker:

instructions, he'll push back and

Speaker:

say, I wouldn't do it like that.

Speaker:

Keep focused on the fact that you're

Speaker:

building an MVP.

Speaker:

So I wouldn't do that right now.

Speaker:

What about might be so Claude

Speaker:

was sort of the first, you know,

Speaker:

build out the architecture, build

Speaker:

out the project plan, build

Speaker:

out the user stories and so

Speaker:

on. Next step is going

Speaker:

over to this tool called V zero.

Speaker:

It's an AI powered generative

Speaker:

UI tool.

Speaker:

So here you set up a new project

Speaker:

into which I upload

Speaker:

the context which has come from

Speaker:

code.

Speaker:

So Claude's given me this whole page

Speaker:

product requirements document

Speaker:

that goes straight into V zero,

Speaker:

and now I can start, you know, can

Speaker:

you design the

Speaker:

homescreen for me?

Speaker:

I want these sections.

Speaker:

Here is my brand colors.

Speaker:

Here's I want a dark theme.

Speaker:

And without me having to write a

Speaker:

very long prompt because it's

Speaker:

already got the entire payload,

Speaker:

the outcomes, the sidebar

Speaker:

outcomes, the project, upload

Speaker:

document outcomes, the chat screen.

Speaker:

I'm not were even better

Speaker:

than that.

Speaker:

What you can put into V zero is

Speaker:

screenshots of other

Speaker:

apps that you like and say

Speaker:

I want that but in

Speaker:

my project context.

Speaker:

So why do I need to reinvent

Speaker:

the wheel? Code and chat

Speaker:

have got the kind of chat interface

Speaker:

fairly well nailed.

Speaker:

If my product did that, I'd be

Speaker:

happy. Screenshot.

Speaker:

Drop into V0.

Speaker:

Give me like that.

Speaker:

And so as a non developer

Speaker:

or non designer, I should say

Speaker:

at this point, you're now

Speaker:

designing screens that are rapid,

Speaker:

right?

Speaker:

And not only does it do the design,

Speaker:

but it then creates all of the

Speaker:

code, whether that's TypeScript

Speaker:

JavaScript in the back end.

Speaker:

So how much did you spend

Speaker:

on it to build

Speaker:

koala?

Speaker:

So when GPT

Speaker:

launched and they pulled a number

Speaker:

out of thin air of $20

Speaker:

a month and as Sam Altman

Speaker:

has said, we basically pulled a

Speaker:

number of out of thin air.

Speaker:

We're going to charge $20 a month.

Speaker:

Well, that's what everyone else has

Speaker:

had to follow. So clod is $20

Speaker:

a month. I pay $20 a month

Speaker:

for GPT.

Speaker:

And I use them both interchangeably

Speaker:

throughout the day.

Speaker:

Rappler, I think, was $120

Speaker:

for a year's license to call

Speaker:

V0 $20 a month.

Speaker:

Kazu I think I pay maybe 30

Speaker:

or $40 a month and

Speaker:

then I'm paying for the others

Speaker:

because I'm testing them out.

Speaker:

That loveable as well.

Speaker:

But they're all in this order.

Speaker:

A couple of things that are a little

Speaker:

bit interesting.

Speaker:

Just when people are thinking about

Speaker:

pricing models for these things, two

Speaker:

things that I've seen are quite

Speaker:

interesting v zero

Speaker:

and lovable.

Speaker:

You can set up a free account.

Speaker:

But when you have a free account,

Speaker:

your designs are public

Speaker:

and so you pay to go private.

Speaker:

And I remember when I was first

Speaker:

designing some of the UI screens

Speaker:

for Koala, I'm sure no one would

Speaker:

have been interested in what I was

Speaker:

doing. But for me, I was.

Speaker:

I felt compelled to make this thing

Speaker:

private.

Speaker:

I'm not seeing that before.

Speaker:

You could imagine that for

Speaker:

Salesforce or HubSpot, you can use

Speaker:

it for free, but your contact

Speaker:

records are public.

Speaker:

Do you want to go private?

Speaker:

So that was an interesting

Speaker:

dimension.

Speaker:

And so you did all of this from

Speaker:

September to Wednesday launch.

Speaker:

December.

Speaker:

December, yeah.

Speaker:

So it was 12 weeks from having

Speaker:

never written a line of code

Speaker:

to having a production app.

Speaker:

I would say that and this

Speaker:

was all done during the weekends and

Speaker:

a few evenings.

Speaker:

There were three weekends

Speaker:

when I literally broke

Speaker:

down in tears in front of my wife

Speaker:

and I was like, Who

Speaker:

am I kidding?

Speaker:

I don't know what I'm talking about.

Speaker:

I should stick to doing what I'm

Speaker:

good at.

Speaker:

You know, what was I thinking?

Speaker:

Like, there's a reason why

Speaker:

developers get paid a ton of money

Speaker:

is because I this is complicated

Speaker:

and I don't know what I'm doing.

Speaker:

And this is one of the problems

Speaker:

with AI development tools,

Speaker:

is that, yes, it gives you this

Speaker:

acceleration from not to

Speaker:

amateur, but when something

Speaker:

goes wrong, you don't know

Speaker:

what's gone wrong because you don't

Speaker:

actually understand at the start

Speaker:

what you're looking at.

Speaker:

And I was trying to deal with quite

Speaker:

complex user authentication

Speaker:

through a third party and I just

Speaker:

couldn't understand why when certain

Speaker:

people were logging in, they

Speaker:

couldn't see certain things.

Speaker:

And I literally broke down in tears

Speaker:

with my wife going, I don't know

Speaker:

what I'm doing. And she'd say, Go

Speaker:

for a walk or go for a run.

Speaker:

That always helps you listen to a

Speaker:

podcast.

Speaker:

And so I'd put on a Lenny Chatzky

Speaker:

podcast and I would come back

Speaker:

inspired and I got I'll

Speaker:

just try this one thing,

Speaker:

and that one thing always solved

Speaker:

it. And I was like, Yes, I'm back in

Speaker:

the game and then would carry on.

Speaker:

So I finished each weekend on

Speaker:

a high.

Speaker:

But Saturday afternoons were my

Speaker:

my low points.

Speaker:

There's some really interesting

Speaker:

conversations happening on

Speaker:

X or on LinkedIn

Speaker:

where you've got real

Speaker:

developers that actually

Speaker:

know what it takes to build a

Speaker:

production enterprise application

Speaker:

and backend infrastructure and

Speaker:

scaling versus

Speaker:

this whole new breed of

Speaker:

indie hackers or

Speaker:

non-technical founders that are now

Speaker:

spinning up these PCs

Speaker:

and the people that have been doing

Speaker:

this for a long time and actually

Speaker:

know what they're doing.

Speaker:

They're like, You think that's

Speaker:

development, That's not development,

Speaker:

You haven't got a clue.

Speaker:

And the same thing happened

Speaker:

when Canva came out

Speaker:

and real designers who are using

Speaker:

Photoshop or whatever are looking

Speaker:

at people using camera going, that's

Speaker:

not real design.

Speaker:

And I know this trend

Speaker:

will continue, but

Speaker:

the tools will improve.

Speaker:

Like I said, I listen to this Lenny

Speaker:

Rich Jet Ski podcast and

Speaker:

always out of my runs and there's

Speaker:

two episodes or two sort of sound

Speaker:

bites that really stuck with

Speaker:

me when I was going through this

Speaker:

process.

Speaker:

The first and I'll

Speaker:

have to remember his name, he was

Speaker:

the founder of Right, which

Speaker:

became Google Docs, and he's

Speaker:

telling Lenny about the process

Speaker:

of writing original Google

Speaker:

Docs.

Speaker:

And he said, We didn't know what we

Speaker:

were doing, but throughout my

Speaker:

entire career, I've just got to the

Speaker:

edge of what I know and fucked

Speaker:

around by getting to the edge

Speaker:

of what you know and fucking around.

Speaker:

You end up learning stuff.

Speaker:

If you stay within what you're

Speaker:

doing, then you don't learn.

Speaker:

So that was one thing on those

Speaker:

days when I was like, I don't know

Speaker:

what I'm doing, I don't know what

Speaker:

I'm doing. I had his phrase in my

Speaker:

mind like, No, but this is where

Speaker:

learning is happening.

Speaker:

And the second episode that really

Speaker:

stuck with me was a guy called

Speaker:

Nikita Baer, who as

Speaker:

well a number of B2C apps

Speaker:

that have gone viral and be bought

Speaker:

by Facebook matter.

Speaker:

I think 1 or 2 of them,

Speaker:

he said this phrase stuck with me is

Speaker:

like, you know, one of the things no

Speaker:

one ever tells you is the moment you

Speaker:

hit virality, you have to build

Speaker:

the whole thing again because

Speaker:

it isn't built for the scale you

Speaker:

need. And I took the positive

Speaker:

in that. I was like, So what you're

Speaker:

telling me is I don't need to build

Speaker:

an enterprise solution.

Speaker:

I just need to build something that

Speaker:

gets to the first 5 or 10

Speaker:

users and solves that problem.

Speaker:

And if I can get that far

Speaker:

and then prove the value in the use

Speaker:

case, then I can get proper

Speaker:

people that actually know what

Speaker:

they're doing and then build for the

Speaker:

next stage.

Speaker:

So that gave me this kind of license

Speaker:

to like, I'm not purporting to be

Speaker:

an enterprise developer that's going

Speaker:

to build a SOC two compliant,

Speaker:

scalable thing, but I don't need

Speaker:

to build that yet.

Speaker:

I just need to prove that I

Speaker:

can solve a problem for someone and

Speaker:

that they would come and log in.

Speaker:

And if I can prove that, then solve

Speaker:

the next stage later on.

Speaker:

So this is me pushing your LinkedIn

Speaker:

content, but there's probably

Speaker:

a piece of content

Speaker:

a week that you write that

Speaker:

I end up sharing internally.

Speaker:

And this week's piece of content was

Speaker:

around public versus

Speaker:

private data and information

Speaker:

with chat shaped.

Speaker:

So you want to talk about that?

Speaker:

One of the things that I'm

Speaker:

always having to remind myself

Speaker:

as someone that would now consider

Speaker:

myself to be in the AI bubble

Speaker:

is that the majority of people

Speaker:

are not in the AI bubble.

Speaker:

They're just, you know, going to

Speaker:

work doing the work the same way

Speaker:

they have. Maybe

Speaker:

they've asked a couple of questions,

Speaker:

you know, what is or how should I do

Speaker:

that?

Speaker:

But they've not really embedded it

Speaker:

in their work.

Speaker:

And to that extent,

Speaker:

a lot of the companies I speak to

Speaker:

either have not got an AI policy

Speaker:

or they've got quite a

Speaker:

pessimistic and restrictive

Speaker:

AI policy.

Speaker:

You know, you must not use GPT

Speaker:

on your company laptop.

Speaker:

You must now upload company data

Speaker:

to GBG.

Speaker:

And definitely because

Speaker:

I've come across almost no companies

Speaker:

that have this, they have not got a

Speaker:

team or enterprise account for

Speaker:

GPT or Clod.

Speaker:

So if you're asking people how are

Speaker:

you using it, I say either I've got

Speaker:

a free account or I pay for the pro

Speaker:

account on my own personal

Speaker:

card.

Speaker:

Now, this is where the data

Speaker:

privacy issues come in.

Speaker:

So if you're on a free

Speaker:

or a paid personal

Speaker:

account for GPT

Speaker:

by default chat

Speaker:

GPT open, I can

Speaker:

use what you upload in

Speaker:

terms of your content.

Speaker:

So that can either be what you write

Speaker:

in the chat or any files that you

Speaker:

upload that is defined as content.

Speaker:

They can use that to train the model

Speaker:

by default.

Speaker:

You can go into the settings

Speaker:

and that is a

Speaker:

I love the way they phrased it.

Speaker:

The setting does not say train our

Speaker:

models.

Speaker:

The setting is would you like to

Speaker:

help improve the model for everyone?

Speaker:

It's a really nice one.

Speaker:

I would like to help improve the

Speaker:

model. It doesn't say what that

Speaker:

actually means.

Speaker:

Your data then goes into

Speaker:

training the model by name.

Speaker:

You can turn that off.

Speaker:

But that is you have to opt out

Speaker:

of that and out of the people

Speaker:

that I speak to, No one does

Speaker:

that because I haven't thought about

Speaker:

it. Most people are low tech users

Speaker:

of it. They just start using it and

Speaker:

uploading things on the teams.

Speaker:

This is GPT specifically when

Speaker:

I say teams or enterprise, but the

Speaker:

same thing on code as well.

Speaker:

If you are on one of those accounts

Speaker:

and by default, your

Speaker:

content is not used

Speaker:

to train the model unless you

Speaker:

explicitly opt in as part of

Speaker:

their feedback model.

Speaker:

So where I find this is quite

Speaker:

amusing in a way that the

Speaker:

companies that are most risk averse

Speaker:

that say, you know, you must not use

Speaker:

AI and we're certainly not going to

Speaker:

pay for an account for you to have

Speaker:

actually push everything

Speaker:

under the radar and under the carpet

Speaker:

to people that are still doing the

Speaker:

work.

Speaker:

But they're using a personal account

Speaker:

which by default gets all of your

Speaker:

company data up into

Speaker:

these models.

Speaker:

And so this is no

Speaker:

time to be an ostrich with

Speaker:

your head in the sand.

Speaker:

La la la la la.

Speaker:

Nothing's happening like it's

Speaker:

happening. People are

Speaker:

using this and you

Speaker:

can choose to ignore it, in which

Speaker:

case you create a data risk

Speaker:

or you can choose to embrace it.

Speaker:

And actually the default

Speaker:

is that your data is not going to

Speaker:

be used to train the models.

Speaker:

So the question of that

Speaker:

and this idea that, hey,

Speaker:

I'm a CEO and I'm

Speaker:

coming into this organization,

Speaker:

venture capital, back to 250 people,

Speaker:

and I want to have a real I

Speaker:

push in this organization

Speaker:

holistically across the company.

Speaker:

So the question to you as a CIO,

Speaker:

is it? What is it that I should do

Speaker:

or what should I think about if I

Speaker:

really want to transform this

Speaker:

organization?

Speaker:

Some of the most common first

Speaker:

use cases when you talk

Speaker:

to companies about using AI.

Speaker:

It comes down to sort of

Speaker:

productivity efficiency.

Speaker:

And there are two ways that you

Speaker:

can approach that.

Speaker:

So if you imagine you've got a 5000

Speaker:

person company and you can say, if

Speaker:

we use AI, we could be way

Speaker:

more efficient.

Speaker:

And therefore, instead of having

Speaker:

5000 people, we could

Speaker:

have 2000 people.

Speaker:

And so that's quite a defensive

Speaker:

cost saving increase.

Speaker:

Our margins way of looking at it.

Speaker:

The other is to say we're 5000

Speaker:

people. But actually, if we used

Speaker:

AI, we can immediately act

Speaker:

like a company of 50,000

Speaker:

people by giving everyone that we've

Speaker:

already got just way

Speaker:

more capability.

Speaker:

Benioff was talking about there not

Speaker:

hiring any more software

Speaker:

engineers this year

Speaker:

because everyone has a sort of wry

Speaker:

smile. Benioff The great marketer,

Speaker:

by his point is we don't need

Speaker:

to hire any more engineers

Speaker:

because we can get the engineers.

Speaker:

We've got to be way more productive.

Speaker:

And you're seeing that across so

Speaker:

many companies.

Speaker:

So the first thing that I'd be

Speaker:

canceling any senior executive

Speaker:

is to say, right, just imagine

Speaker:

that you could supercharge the

Speaker:

people that you've got and

Speaker:

get them to be way more capable

Speaker:

to be able to work in different

Speaker:

countries in different domains, to

Speaker:

know what way more than they do

Speaker:

today.

Speaker:

How would you approach that?

Speaker:

And a lot of this comes back to

Speaker:

not trying to do anything overly

Speaker:

complex. So even a couple

Speaker:

of years ago, when you talk about

Speaker:

AI, people are, you know, we're

Speaker:

going to use AI to completely

Speaker:

disrupt our supply chain

Speaker:

or we're going to use AI

Speaker:

to completely rebuild our

Speaker:

forecasting of footfall in a retail

Speaker:

environment or something.

Speaker:

It's quite a complex thing.

Speaker:

Machine learning, data science.

Speaker:

The big win right now is just get

Speaker:

every single person that works

Speaker:

in h.R.

Speaker:

Is a financial analyst.

Speaker:

That is a marketer that works in

Speaker:

customer support.

Speaker:

They have work that they have to do

Speaker:

as a human being.

Speaker:

And i sort of split this down into

Speaker:

three buckets.

Speaker:

There's there's work that you

Speaker:

receive from other people, so

Speaker:

someone gives you some requirements.

Speaker:

A customer gives you a request for

Speaker:

proposal.

Speaker:

Someone gives you a quote

Speaker:

or a contract.

Speaker:

Maybe there's a new regulation or

Speaker:

new compliance rules.

Speaker:

There's something that someone else

Speaker:

has given you.

Speaker:

And as a human being, you have to

Speaker:

understand that, analyze it,

Speaker:

summarize it, work out, what

Speaker:

are the most important bits and the

Speaker:

risks.

Speaker:

That's work that's given to you.

Speaker:

You've then got the work that you

Speaker:

actually do as a person.

Speaker:

So, you know, I need to

Speaker:

research accounts.

Speaker:

I need to prepare a proposal.

Speaker:

I need to manage a direct report.

Speaker:

Those are the things that you have

Speaker:

to do yourself.

Speaker:

And then there's work that you pass

Speaker:

on to other people.

Speaker:

So I need to create a proposal.

Speaker:

I need to create some requirements.

Speaker:

I need to create a summarization

Speaker:

for my manager.

Speaker:

I need to create an investor

Speaker:

briefing, whatever that might be.

Speaker:

And so these are very human things

Speaker:

that you have to do.

Speaker:

And how can you bolt on

Speaker:

the skills where they are to be able

Speaker:

to accelerate that much

Speaker:

more capable at ingesting

Speaker:

information and new requirements,

Speaker:

much more effective at doing the job

Speaker:

that I've got to do and a much

Speaker:

higher quality of output and

Speaker:

speed of output that I'm delivering

Speaker:

to other people.

Speaker:

And in it, you talk

Speaker:

about the ten x developer or the ten

Speaker:

x engineer, you know, what is the

Speaker:

ten x h.r.

Speaker:

Admin. What's that?

Speaker:

Ten x customer support

Speaker:

representative.

Speaker:

There's people that are kind of

Speaker:

biochemically enhancing themselves

Speaker:

with these tools suddenly

Speaker:

way more capable than the person

Speaker:

that sat on the desk right next to

Speaker:

them.

Speaker:

So I think that's a great vision.

Speaker:

There are a lot of people where all

Speaker:

CEOs, we're like, Yeah, okay,

Speaker:

whatever vision, how do we do it?

Speaker:

And dropping it down a level, I can

Speaker:

explain some of the ways that I use

Speaker:

cloud or chat type of ended

Speaker:

up buying Claud instead.

Speaker:

Because of the ability to

Speaker:

mimic your writing, I find that

Speaker:

Claude creates my writing better

Speaker:

than chatbot, creates my writing out

Speaker:

of the box, so that's why I

Speaker:

subscribe to it.

Speaker:

I use it for things

Speaker:

when I have to be creative and I'm

Speaker:

not being very creative.

Speaker:

So I want an image

Speaker:

for a presentation that evokes

Speaker:

a certain feeling.

Speaker:

And in the old world,

Speaker:

I used to have an image and I

Speaker:

couldn't find it. And I talked to

Speaker:

our designer and he has the entire

Speaker:

Adobe back catalog and he

Speaker:

gets me some photo and it's done.

Speaker:

But now I can

Speaker:

either describe an image and have it

Speaker:

generated or I can just say

Speaker:

I don't really know what image I

Speaker:

want for

Speaker:

exceeding our target this quarter.

Speaker:

You know, like I can only come up

Speaker:

with like some really generic ones

Speaker:

and I'll say, give me ideas of

Speaker:

images that'll work and I'll come up

Speaker:

with loads of images. And I'm like,

Speaker:

No, I don't like that one.

Speaker:

I do like this one.

Speaker:

Okay, yeah, that sounds good.

Speaker:

Edit it this way.

Speaker:

And now create.

Speaker:

And then Claude doesn't I?

Speaker:

She creates images, but Chatty

Speaker:

Betty does.

Speaker:

Or Dolly does.

Speaker:

And so then I'll say. Write a prompt

Speaker:

for Dolly that will give me the

Speaker:

image. And then I'll put it in,

Speaker:

create the image, and then edit

Speaker:

between the two.

Speaker:

And so and and all of that can be

Speaker:

way faster than it used to take me

Speaker:

to trawl through images to try

Speaker:

and evoke the sensation

Speaker:

that I want to evoke for customers.

Speaker:

Or I need to write the weekly

Speaker:

report or the weekly customer

Speaker:

update. And there's bits of it that

Speaker:

are written for me in bits that are

Speaker:

end, and I have writer's

Speaker:

block. And I'll just say I

Speaker:

want to talk at the end of the year,

Speaker:

end of the week, I've actually

Speaker:

created a project for this.

Speaker:

It's the end of the week.

Speaker:

I need to make people feel this way.

Speaker:

These are some topics that I'd like

Speaker:

to talk about. Everything I'm

Speaker:

writing is really stupid.

Speaker:

Write me a paragraph.

Speaker:

That's not shit.

Speaker:

Obviously it's a better prompt than

Speaker:

that, although sometimes that is the

Speaker:

prompt and it'll at least give me

Speaker:

something I rarely cut and paste and

Speaker:

use that itself.

Speaker:

But it's a way of like

Speaker:

whenever I'm stuck with

Speaker:

the too hard project, that is me

Speaker:

having to think.

Speaker:

I turn to Claude to help me think.

Speaker:

And oftentimes it just releases

Speaker:

the writer's block in me

Speaker:

to produce something.

Speaker:

And that's not an efficiency thing.

Speaker:

That's actually like a thinking

Speaker:

thing. I mean, it is efficiency, but

Speaker:

it's not like summarizing my emails.

Speaker:

I'll give you a specific example,

Speaker:

which for SEO

Speaker:

is a really sort of tangible

Speaker:

way of using this.

Speaker:

Anyone that's in a sea level role

Speaker:

has got direct reports

Speaker:

managing or leading those direct

Speaker:

reports. Probably when you became a

Speaker:

leader, you had these great

Speaker:

aspirations of how you wanted to

Speaker:

be, you know, a great leader and

Speaker:

a coach and advisor.

Speaker:

And too often we get dragged down

Speaker:

to just being a manager,

Speaker:

a boss, and you're just going

Speaker:

through the day to day

Speaker:

exercise, take a code

Speaker:

project or

Speaker:

just launched projects just before

Speaker:

Christmas.

Speaker:

Create a project one for

Speaker:

each of your direct reports.

Speaker:

So I create a project for

Speaker:

Bethany As and

Speaker:

in that project I can upload

Speaker:

certain documents.

Speaker:

Now, I would add in

Speaker:

my regular 1 to 1 notes,

Speaker:

I would add in maybe the performance

Speaker:

review that came out of Workday last

Speaker:

year. I might add in some

Speaker:

chaos that have been set.

Speaker:

Maybe there was a desk review

Speaker:

assessment or something like that

Speaker:

and maybe just do my chat.

Speaker:

So my knowledge of working with you,

Speaker:

but I understand a little bit about

Speaker:

your sort of personal career

Speaker:

aspirations or maybe personal

Speaker:

goals around houses

Speaker:

or family or whatever that might be.

Speaker:

And having done that,

Speaker:

you can then start to ask

Speaker:

this project about

Speaker:

those kind of creative things.

Speaker:

You know, I'm preparing for my 1 to

Speaker:

1 with Beth.

Speaker:

Can you help me think through some

Speaker:

stretch goals that would support

Speaker:

what best career aspirations are?

Speaker:

And I need to provide Beth

Speaker:

as some challenging feedback around

Speaker:

the way that she handles a recent

Speaker:

project.

Speaker:

Can you help role play with me?

Speaker:

How I could deliver that feedback to

Speaker:

Beth? Understanding Desk and

Speaker:

the way that she is going to respond

Speaker:

to that.

Speaker:

Suddenly you've got ten

Speaker:

individual leadership co-chairs

Speaker:

for each of your direct reports,

Speaker:

and you know, that's useful for

Speaker:

managers, but it's about

Speaker:

thinking about what's the work that

Speaker:

someone is already doing.

Speaker:

They have to do it.

Speaker:

And how do we use AI

Speaker:

to support me and being better

Speaker:

at that specific role?

Speaker:

It might be managing someone.

Speaker:

You might have a project for a

Speaker:

customer that you're selling or

Speaker:

supporting into.

Speaker:

It might be a project for a country

Speaker:

that you're looking to open up.

Speaker:

It might be a project, in my

Speaker:

case, with a product that

Speaker:

I'm trying to build.

Speaker:

So it's thinking about breaking

Speaker:

down. What are the bits of work that

Speaker:

you're already doing and how do

Speaker:

we give enough context

Speaker:

around that to be able to give me

Speaker:

what I want?

Speaker:

Can you give us one example?

Speaker:

I need to ingest a wide swath

Speaker:

of information, either summarize

Speaker:

it or pull out some takeaways in

Speaker:

terms of something that I'm trying

Speaker:

to analyze or what have you.

Speaker:

I'm going to mention a tool called

Speaker:

Notebook Alarm, which

Speaker:

went pretty viral in the autumn of

Speaker:

2024. It comes from Google, so you

Speaker:

can just go to notebook alarm dot

Speaker:

Google.com.

Speaker:

And what they've done is take a very

Speaker:

sort of intuitive approach to using

Speaker:

AI. So you can upload

Speaker:

a source and a source

Speaker:

could be a document, it could be a

Speaker:

YouTube video, it could be

Speaker:

a website.

Speaker:

And what notebook will do

Speaker:

is ingest that source

Speaker:

and use that and only that

Speaker:

to drive a set of AI

Speaker:

outputs and responses.

Speaker:

So it's not accessing the wider

Speaker:

Internet.

Speaker:

And a couple of things that it

Speaker:

creates off the back of that

Speaker:

document is a briefing

Speaker:

document.

Speaker:

It creates some fake news

Speaker:

and the to resistance,

Speaker:

it creates a podcast

Speaker:

interview between two

Speaker:

AI generated hosts

Speaker:

and an example

Speaker:

that a provider may.

Speaker:

A little link to some of the

Speaker:

recording of this.

Speaker:

But is the the new EU

Speaker:

Air Act, which

Speaker:

starts to come into force in

Speaker:

February 2025.

Speaker:

So in just a month or so from now?

Speaker:

Now the EU Air Act is 144

Speaker:

page document, very dry

Speaker:

as you'd expect, and difficult for

Speaker:

most people, if not legal, to

Speaker:

quickly figure out.

Speaker:

Does this relate to me?

Speaker:

How must I consider it?

Speaker:

So what you're able to do in this

Speaker:

example is upload that

Speaker:

document and then suddenly

Speaker:

it generated for me a 31 minute

Speaker:

podcast interview with these two

Speaker:

hosts discussing it.

Speaker:

But even better than that, you

Speaker:

can join that podcast.

Speaker:

So as you listen to it, you can say,

Speaker:

Sorry, I don't understand that, or

Speaker:

just help me to understand that.

Speaker:

So suddenly you're having a

Speaker:

conversation about

Speaker:

this complex subject.

Speaker:

Right, Right. So it's it's right.

Speaker:

You can interrupt the podcast hosts

Speaker:

and basically ask a question

Speaker:

midstream. You're like, Hey, podcast

Speaker:

host, I have a question

Speaker:

100%.

Speaker:

And they go, Hey, what's up?

Speaker:

You know, what's your question?

Speaker:

And then you also you say, you know,

Speaker:

you talked about the

Speaker:

air coming in, but like, what's

Speaker:

the first deadline that

Speaker:

I have to be considering?

Speaker:

And they go, you know, well,

Speaker:

February the 2nd, 2025.

Speaker:

And so you're having this

Speaker:

conversation now.

Speaker:

This is about the Air

Speaker:

Act. But think about all of the

Speaker:

things that you have to consume.

Speaker:

Maybe it's the Mssa that's been sent

Speaker:

to you from a supplier.

Speaker:

Can I give you an example of what

Speaker:

we're looking to do?

Speaker:

So we're also it's

Speaker:

interesting that you've mentioned it

Speaker:

because my first experience

Speaker:

with it was yesterday when one

Speaker:

of our Gen I people pulled

Speaker:

together information

Speaker:

on ideas of what we're going

Speaker:

to do as a podcast.

Speaker:

Not only does it do it as a podcast,

Speaker:

but it can also play on your voice.

Speaker:

So it's his voice doing

Speaker:

the podcast.

Speaker:

And what we're going to experiment

Speaker:

is, do you remember years ago,

Speaker:

I can't remember which company it

Speaker:

was. It was a company that I worked

Speaker:

with that would give all of their

Speaker:

sales team CD's

Speaker:

to listen to for sales

Speaker:

enablement while they were on the

Speaker:

road. It was either Microsoft or

Speaker:

IBM, I don't remember.

Speaker:

And so what we're going to

Speaker:

experiment with is taking all

Speaker:

of our sales enablement content

Speaker:

and turning that into

Speaker:

an internal podcast for our sales

Speaker:

team so that while they're on

Speaker:

the road, not that people are

Speaker:

on the road as much as possible can

Speaker:

consume that content.

Speaker:

And we're looking at a couple of

Speaker:

tools to basically create

Speaker:

an internal playlist where you

Speaker:

can find all the content as a

Speaker:

podcast.

Speaker:

But I didn't realize about the

Speaker:

interactivity, so that's super cool.

Speaker:

So yeah, it's only just released

Speaker:

maybe a week or so ago that they put

Speaker:

the interactivity in.

Speaker:

It's amazing.

Speaker:

There's a maybe a two second

Speaker:

delay after your question

Speaker:

wallet reconfigures,

Speaker:

but it's amazing.

Speaker:

So my one question is

Speaker:

Gemini.

Speaker:

It was shit have heard that Gemini

Speaker:

two is not shit.

Speaker:

Have not played with it yet.

Speaker:

We're going to experiment.

Speaker:

Worth it or not.

Speaker:

So when I'm talking to clients,

Speaker:

I talk about three things.

Speaker:

I talk about Gemini chat

Speaker:

and Cod as being those two

Speaker:

providers that people should be

Speaker:

thinking about and definitely

Speaker:

want to be watching, especially if

Speaker:

you're a if you're a Google

Speaker:

workspace customer, then

Speaker:

that has to make sense.

Speaker:

I just say on that, it's one of

Speaker:

these things that is so different

Speaker:

to SAS five years ago.

Speaker:

If you're buying a CRM system or

Speaker:

you're buying an ERP system, you

Speaker:

go out to market, you pick which one

Speaker:

and you pick one.

Speaker:

That is not the way that it works

Speaker:

here. People are not going out to

Speaker:

market to pick one Al-Alam.

Speaker:

You will have this composable

Speaker:

architecture.

Speaker:

You will have multiple,

Speaker:

maybe some closed source, maybe some

Speaker:

open source, and you'll use

Speaker:

different ones even within the same

Speaker:

workflow.

Speaker:

You know, this is doing the research

Speaker:

and this is doing the writing.

Speaker:

And so from a CEO perspective,

Speaker:

from a CIO or CTO perspective,

Speaker:

I think about Lego blocks.

Speaker:

What you're building is a box of

Speaker:

Lego that people can build

Speaker:

what they want at the time.

Speaker:

This isn't like a traditional

Speaker:

procurement, so you would have

Speaker:

Gemini plus the others in

Speaker:

your box of Lego.

Speaker:

If our listeners can

Speaker:

only take away one thing from

Speaker:

today's conversation, what is

Speaker:

it?

Speaker:

It is to plant

Speaker:

the seed of

Speaker:

use and every

Speaker:

single one of your business

Speaker:

teams.

Speaker:

I say business to me.

Speaker:

This is not an IT project.

Speaker:

So for every one of your

Speaker:

teams, how can you give them just

Speaker:

one way of using

Speaker:

publicly available tools

Speaker:

called Gemini to improve

Speaker:

one aspect of their work?

Speaker:

Because it is those people

Speaker:

that are closest to the business

Speaker:

problem and the business process.

Speaker:

And if you can show them one

Speaker:

way of using this, that is where

Speaker:

you will suddenly find all of

Speaker:

these new ideas come from.

Speaker:

So figure out how to get this

Speaker:

groundswell going at the

Speaker:

core of your business and bring

Speaker:

that back up to the top rather than

Speaker:

going top down.

Speaker:

Lovely. Thank you, Charlie Cohen,

Speaker:

for joining us on the operations

Speaker:

room.

Speaker:

If you like what you hear, please

Speaker:

subscribe or leave us a comment.

Speaker:

And we will see you next week.

Next Episode All Episodes Previous Episode

Listen for free

Show artwork for The Operations Room: A Podcast for COO’s

About the Podcast

The Operations Room: A Podcast for COO’s
We are the COO coaches to help you successfully scale in this new world where efficiency is as important as growth. Remember when valuations were 3-10x ARR and money wasn’t free? We do. Each week we share our experiences and bring in scale up experts and operational leaders to help you navigate both the burning operational issues and the larger existential challenges. Beth Ayers is the former COO of Peak AI, NewVoiceMedia and Codilty and has helped raise over $200m from top funds - Softbank, Bessemer, TCV, MCC, Notion and Oxx. Brandon Mensinga is the former COO of Signal AI and Trint.

About your host

Profile picture for Brandon Mensinga

Brandon Mensinga