And I won't tolerate anyone inside my organization coming to me with a plan without first having talked to AI.

Now what I do is I've got a monitor, it's got 1 browser and three tabs open.

Perplexity, GPT and Claude, which are my 3 go to a is we're 25% more productive across the company.

It's as simple as that.

We ask AI to play the role of a board member.

Claude has been the best performing and performing at 90% of the humans.

Greg, welcome to the show.

Thank you.

Good to be here.

Yeah, super excited to do this.

So I was taking a look at your background just earlier.

You you've basically done it all like you worked at a big company, you've started companies, looks like you spent some time in venture capital and you know today you're back as CEO again and you're CEO of a company called Section.

What is Section?

Maybe.

Maybe tell everybody what it is.

Sure.

Well, first of all, talk about me for a second.

In terms of my career, mostly as an entrepreneur.

And I would say a a more typical career, which is no unicorns, no generational companies, no billion dollar access, which my kids keep reminding me of.

I've had a good mix of, you know, flame outs, zeros, singles, double s, maybe no home run yet, but you know, companies have sold in total for $250 million in terms of exit value.

So not, you know, not nothing.

Not too shabby.

Not too.

Shabby, but but nothing that Silicon Valley gets excited about.

Or my kids.

But the best one is the one you're working on now.

So, so this one, I think we'll get next to it for sure.

And you know, we'll see how big an exit.

And I'm, I'm a hired gun CEO in this company.

I, I did not found it.

So I have a more typical, you know, CEO stake if you're a hired CEO and we are and we've pivoted by the way section pivoted about two years ago and we could talk about pivots if we have time.

We are an education company that has pivoted to an AI enablement company.

So we were teaching managers and leaders online for the first three years of the company's history in terms of strategy, storytelling, financial analysis.

We thought think about us as originally as an executive Ed online executive Ed for tech executives primarily.

I played with ChatGPT +2 February the 1st 2023 and that day I pivoted the company.

I said to myself, and by the way, the only way I've made money in Silicon Valley is catching a wave.

The best way to have success is to find and when I played with ChatGPT Plus on February the first over 2 years ago, I realized this was going to be the growth.

And so I immediately said to myself and to the rest of the company, let's become an AI school and, and teach others how to use AI and then teach organizations how to use AI because, you know, this is it.

This is going to be the kind of transformational technology.

So we did that two years ago.

For the first six months, everybody and my own company ignored me and thought I was a typical CEO going crazy.

And I was like slacking and texting and hey, you're, you know, read this, try this, check this out.

And everybody said like, Hey, we got day jobs, we're too busy, you know, kind of go away.

So, you know, I eventually realized I, I needed to be serious myself about change management inside of my own company.

Kept, I kept putting stakes in the ground.

We're going to, we're going to teach this class by this date, we're going to start to shift our new customers into our AI Academy.

We really just, if you pivot, you got to pivot and not go back.

And so it took me a while to realize my team thought I might change my mind and I wasn't that serious about this AI stuff.

I've got to remember this was two years ago, so it probably looked less, less inevitable 2 years ago even than than looks today.

Eventually after six months, we're like, OK, we're all in.

We're not going back.

We will be the AI Academy and we will help organizations make this transition to get their workforces to become AI powered.

So that's what we've been doing for the last two years.

Yeah, I, I love it.

And by the way, you're, you're not the first person to tell me a similar story around, you know, I was very early started telling my teams, trying to get them.

And they're like, Oh yeah, here it goes.

The CEO finding the next shiny object and and thinking it, you know, the world has changed.

But like this time it really did.

So yeah, good on you for.

And that's a quick pivot.

I mean, I remember checking out Chachi PT early, but it didn't occur to me that we should pivot the company until much further down the line.

So yeah, good on you for for noticing that early.

So the executive education stuff is really important and useful, right?

I mean, so if I have this correctly, so if I'm ACEO of a company and what I want is for my people to be more AI forward, more AI educated, like they come to section and and there's just course material and and everything.

Yeah.

And they can take advantage of our kind of our public offering, or we'll build custom courses for them and custom change management programs.

And really our business is becoming more change management and what we call workflow redesigns.

If you really think about how AI is going to land and then stick inside the organization, it's the training is a piece of it, but actually a pretty small piece of this piece of it.

And this is much more about change management.

We have so many cultural beliefs about work.

We have so much anxiety about AI.

We have so much skepticism about AI amongst our workforces that we've got to get through that.

And and by the way, most organizations aren't, I would characterize most organizations as failing in their deployments of an LLM or of a general purpose AI, something like ChatGPT for Enterprise Oregon, Gemini Pro or Microsoft, you know, Copilot.

I would say most organizations and certainly mid to larger ones are absolutely stalling out because they haven't really acknowledged and appreciated the, the, the amount of change management and behavioral change required here to get employees to use AI.

And by the way, most employees think that they're training their replacements.

And so, you know, and Cos are not being upfront and talking enough about the impact that AI will have on our jobs.

And, and that doesn't mean necessary layoffs.

It will mean layoffs, but it will also mean, you know, how our jobs change and and how the tasks, some of them get outsourced to AI and some of them stay with the human.

I mean, there's a lot going on here.

This is not software.

I think what really the failure of corporate, corporate America so far with AI is thinking that it's like software and, and deploying it like software.

So, you know, buying it first of all and signing a three-year agreement, a license agreement with, you know, Microsoft or, or open AI fine.

But then deploying it kind of like it's a new ERP or a new CRM or a new marking automation software, a new productivity suite.

It's not, it's nothing like that.

It's this crazy powerful cointelligence that behaves unpredictably.

It's not predictable, it's unreliable.

It's good at so many things, but not great at one thing yet, right?

Software, the stuff we use everyday at work, is typically good at one or two things and reliably very good at it.

Yeah, that's been the playbook, right?

Like build something that's really good at one thing one.

Thing very narrow right most and this is the opposite.

This is this thing is can code and create images and write and be a thought partner and help you figure out product Rd.

maps.

I mean, this thing can do almost anything, sometimes really well and sometimes quite poorly.

And when you imagine, you know, if you told an employee use this software and some days it's going to work and some days it's not going to work.

Sounds like a human.

Why would I do that?

Like, it's not.

That's not helpful anyway.

So there's a, you know, we're early.

We're two years in.

Here's what we do know though, right?

People love it.

ChatGPT will get to a billion active users sometime this summer, right?

Sam Allman, let's slip at Ted a few weeks ago that they're probably at 6700 million active users.

That's free.

That's the free version, right?

They're probably at around 30 to 40 million paid users.

But still, the fact that this technology as unreliable as it is and it's hard to use really as it is in some ways will get to a billion users sometime this summer is, is stunning.

And that's it's a consumer and small business revolution.

It's not an enterprise.

Enterprise always lags.

I think with AI, they're lagging even further than they typically do, meaning that they're slower to act and they, and they do have all these kind of deployment challenges.

That's where our business has become really from upscaling.

It's, it's really moving into more building these programs, more holistic programs for companies over the course of 12 to 36 months, where we really come in and help them get AI to stick inside their organization, which is changing these behaviors and getting AI into workflows and, and really changing how work gets done, you know, with these technologies.

So this is interesting.

I mean we, we were chatting, you know, before we hit record.

I mean you've seen a lot of these deployments.

So, so the question is when it has worked well and I'm sure there are some success stories of companies who who've done it well.

You know, what do they do?

Like the ones that do actually, you know, succeed at this, You know, what do they typically do that others don't like?

What causes failure or what causes success?

Sure.

I'd say the first thing they do is they figure out their AI policy, especially as it relates to data and security and so on.

But they also figure out their point of view.

So you're not going to have success if all you deploy is an AI policy because all that does is scare people and tell them what not to do.

So you need to deploy that.

And that's kind of, you know, protect your downside, so to speak.

But as a leadership team, you've got to deploy your AI manifesto your AI point of view.

What do you think about AI?

Are you encouraging its usage?

Do you think it's working smart or is it cheating?

A lot of people think it's cheating to use AI.

So then no one's going to do it.

So you've got to really as an exec team, be much more forward, you know, and outspoken around.

You expect people to use it.

You think it's working smart.

You think that you want ideas that have been stress tested with I, with AI before, you know, they're presented to you and so on.

So listen, we've seen in the last couple weeks, I'm sure you've seen, you know, the memos from Toby Latke at Shopify or the CEO Duolingo, and then now the CEO of Fiverr.

You can see what Toby at Shopify, you can, you can kind of imagine what happened.

He got frustrated.

He got frustrated because he's using AI everyday, getting a lot of value from it.

And then he's looking across his organization, big organization, 8000 employees, and he's not seeing the levels of adoption.

So to his credit, I think, you know, he just put a stake in the ground and a bit of a carrot in a stick.

They care of being, hey, you know, we're going to get value from AI.

We're going to be able to, you know, outsource some of this work to AI.

We're going to be able to focus on adding more value to our projects into our jobs.

And also the stick, which is it's going to be in the performance review every year.

Your AI proficiency is you're going to be measured on it and you're not going to be able to get human headcount approved unless you can prove to us that the AI can't do the work.

So I, I think something like that's appropriate, which is, you know, you can go all carrot, you can go all stick.

It's probably a mix, but you have to have a point of view.

That's the first thing.

Second thing, you have to give them a good AI, pay for a good AIA.

Good AI means latest version of the models.

What we see when we survey organizations, they might have deployed Copilot, the free version of Copilot, which is not that useful.

And they'll have as much usage of GPT, ChatGPT, unsanctioned.

People aren't stupid.

They want the best tools.

And so you're always going to have some part of your employee base or your workforce.

What are early adopters?

They are the curious.

They are the, you know, those that are ambitious and they're going to find the best tool and use it whether you approve it or not.

So just realize that's what people want.

They want the best AI and you should pay for it as the employer.

You know, why are you letting people, you know, letting people pay for it themselves?

Like don't be cheap.

So, so it's true and I don't know if you would agree with this, but like part of me sometimes, you know, I'm using these tools and say chat cheap T and it's like so good.

The amount of value that it gives me personally is so mind blowing that I'm like, there's no way this is going to stay $20.00 a month.

Like once once like, you know, it becomes well known like the actual value this stuff delivers, I feel like the it's going to cost a lot more.

You know, counterintuitively, maybe people will say that no, you're wrong.

Technology costs less over the course of time.

But yeah, I mean, there is this like element of, well, at some point you're going to capture more and more of the value you deliver.

I mean, we, we should talk about that a bit because I think it's pretty fascinating because I think I think we're, you're right or both arguments are right.

In some ways.

The cost of AI is coming down in part because these models seem to be converging in terms of their base capabilities.

And I think open AI probably looks back and says, wow, we, we might have screwed up here, setting kind of the floor in the ceiling at $20.

And the Enterprise Products cost a little bit less, but not much, right?

ChatGPT for teams is 25, but they're basically, you know, close to free and you get a lot of value.

You're right.

And they, they've had a couple of premium products, like 200 bucks a month.

But the bottom line is most of us think that amazing AI cost 20 bucks a month.

So I, I think in some ways they've done the industry a huge disservice, but by setting the price so low.

And that's why they're saying things like we want to charge for agents in a completely different way.

Essentially as a percentage of the human when opening eyes said our agents will cost between 2000 and 20,000 bucks a month.

What they're saying is we're going to try a new sales pitch here in a new pricing model, which is the human costs 100 or 150, whatever the number is.

And if your agent is doing some fraction of what the humans doing, maybe all of that job, you, you know, you, you should pay a lot more for the agent.

I don't know if that works though, because again, other side of that coin is it's still just software.

Agents are just software that are trained up.

And maybe there's downward or deflationary pressure on those prices, But you can see what's starting to happen.

The AI companies want new and you know, they want a Better Business model than 20 bucks a month.

And, and I think agents are the way they think they might find, find some new pricing, which could be, you're right, 1000 or, or, or several thousand a month, because these things are so performant, These technologies are kind of mind blowing, right?

An hour, an hour of analyst time converts to a minute of reasoning time that that's what deep research does.

If you are a heavy deep research user, you'll know that the quality is pretty much like a, an MBA level analyst or you know, a research analyst, a health researcher.

Health researchers, because I've hired one recently, cost 100 to $200.00 an hour for someone who's an experienced researcher in medicine or health.

And so you kind of get that for for a dollar, you know, you got a minute of reasoning time for an hour of human labour.

It's pretty crazy.

And I have to say also the other thing I love about all the deep research models is the just like the slight UI tweak of they could have said loading, but just by saying thinking, I'm like, yeah, think longer.

And now I actually get upset.

It's like, you know, if it says thought for four seconds, I'm like, hey, don't cheap out on me here.

Go work harder.

Go think longer.

I want a better answer.

Like what a, what a brilliant like user interface for that Super cool.

So we, we talked about, you know, how, you know, companies that are doing this well and, and, and those that are not.

I know that since you guys pivoted the company, obviously you're using AIA lot.

What's cool about the role that you're in is a, you get to work with a lot of companies.

You see what, what's at the edge, who's succeeding, who's failing?

And I'm sure along the way you have lots of great ideas about, you know, ways to bring it into section and the way that you guys use AI everyday.

I'd love to know, like how would you say that AI has impacted the work at section?

Sure.

We're 25% more productive across the company.

It's as simple as that.

It took a year and we've gained 25% productivity.

Now listen, if we couldn't, I think shame on us.

Meaning we're in the, we're in a content business essentially.

I mean, we're, we're, we're a bunch of smart people creating plans, creating lessons, creating case studies, you know, learning and change management programs.

But it's all basically human IP and and then the output is content in some form.

So we are Ground Zero, Ground Zero for AI.

Like if we couldn't take advantage of journey of AI, then you know that that was on us, not, not the technology.

Same for frankly any software company.

So if you're a language intensive with a lot of human capital and a lot of IP basically going into your product or service, you absolutely should be able to deploy these technologies.

And within 612 months, see I think quite significant productivity gains.

And when I told the board at the beginning of this year, as we would see a, a 25% another increase in our productivity this year by just using AI everywhere.

And and that's absolutely what we're seeing.

We're growing.

So we're still hiring, but if we weren't growing, if we were not growing as an organization, we absolutely would not be back filling positions and you know, we might be reducing headcount.

That is the reality of of AI certainly in industries or companies where you can really take advantage of it, which is what I call language intensive businesses with high human capital inputs.

So again, we're growing.

So we're just hiring less people and we're making our products faster.

You know, everything we're doing is a little bit faster, a little bit sharper.

I think we've certainly outsourced a lot of grunt work.

So we started there.

I would say, I'd say year one was more about outsourcing grunt work to AI.

For example, using AI to, to process all of our customer surveys, all the qualitative and

N PS:, N PS: scores and qualitative feedback we get.

And all that's done by AI now and then the human gets the gets the first report from AI and then, and then we take action based on what the human thinks.

But all that processing, all that consumer sentiment and surveys is now done by a as an example.

And that was hours of drudge work for someone every month.

So there's probably 50 to 75 AI powered workflows now inside of our small team of 30 people where they IAI is doing a significant amount of that workflow.

Where was human only, you know, a year or two ago.

Now I'd say we're pushing more into changing the product.

So I think about this as optimal 1st and accelerate second.

And so optimize means make everybody more productive and try to, you know, raise your levels of productivity and and maybe reduce your reliance on a new headcount and then begin to change your product or service.

Try to accelerate your business in some way by introducing AI into the product or service in a meaningful way that either raises the price or increases customer loyalty so you have better retention or you know, whatever the metric is that you care most about.

So that's where we are been very focused on the past 12 months and we released a new product two weeks ago called Prof AI that is an example of this, which is it is an AI powered AI coach.

So it's kind of meta, but we're using basically AI to teach people AI, and we're doing that through a chat interface.

So we built an application on top of the Open AI and Anthropic Cloud models.

Oh, super cool.

One thing that was interesting, what you said is you actually had a quantitative measure.

You said productivity's up 25%.

How do you get that number?

And, and, you know, when people are thinking about, OK, we're kind of, you know, messing with AI, few people have ChatGPT accounts.

It's not centralized.

People are doing their own thing.

And, you know, how do you come up with a number?

So if Aceo is going to walk to the board and say we expect this, like how do they come up with what this is?

First of all, it's hard and first of all, it's very context and and company and and workflow specific.

We had some metrics though that we could use in terms of students we serve, enterprise clients, we serve new deals, we sign every month in terms of new business.

So we just took a look at those metrics baseline from two years ago and then we compared to today in terms of how much headcount have we added and how much more are we serving or creating.

So you need a baseline and it's, it's something like what are your outputs?

However you measure your outputs, you know, customers or, or you know product output or codes, you know, lines of code written, whatever might be.

So established a baseline account for growth and then look at your head count.

We're just, we're hiring less people to accommodate double the growth.

We've kind of, you know, triangulated to their, I wouldn't say it's science.

And if ACFO came in and really wanted to, you know, stress test my 25% number, I don't know if it would, you know.

Pass the test of a, you know, of a tough CFO, but that's what we feel like.

That's what it feels like.

And that's what we can see in our hiring.

We've we're, we're, you know, we're growing at 100% a year, but we're not even close to growing the headcount at that rate.

Like we're getting more efficient in terms of what we do in terms of the outputs we create and the customers we serve.

So I think that's the basic way to do it.

It'll show up late.

This is the issue with AI mean you ask a great question.

This is the issue with AIAI right now, meaning enterprise AI, it's hard to measure and the impacts are going to come later and the impacts are likely going to come in reduced hiring.

So that can take 612-1824 months to, to kind of flow through the system, if you will.

And, and by the way, most managers don't volunteer headcount.

If you're a manager, you know, you like headcount, you want more of it, not less.

And so it's really going to have to come later, frankly, from CFOs.

And that's what's going to start to happen if it isn't already happening, which is no, we're going to hold hiring for a while and see if the AI can step into that gap.

And you know, we'll know, but it takes 612 months to know.

You know what's interesting?

I I saw this, one of our friends at Vidyard posted this.

I don't know if you saw this too, but they kind of posted a cost of a pull request and how the cost of a pull request has gone down, you know, the number of them versus the number of engineers that, you know, that were involved.

But, but I like that as a as a metric in that, you know, just choose whatever output metric you have and then figure out some ratio to people.

The other the clever thing, I mean, I love the, you know, Toby talking about, you know, measuring this stuff and, and in performance reviews and and evaluating employees that way.

But another one that's really cool is, you know, I've also heard so Andrew Weightman was just on the show and and he was talking about using just having everybody has to have an OK R related to AI across the company.

And so if it starts to become part of like your objectives, then it really starts baking itself into the organization and also shows, like, the company's serious about this stuff.

Absolutely, Yeah.

Listen, we we do goofy things like AI, AI shout outs.

And when we first started doing them two years ago, it was pretty awkward because, you know, you get all hands meeting always, you know, we do human shout outs.

Everybody does one or two people get shouted out kind of each week.

I said, listen, I want to do AI shout outs.

And again, for the first few months, this is weird.

Like we're going to shout out Claude, shout out GPT, shout out, you know, Gemini.

But it just starts to work in that it tells people, let's acknowledge when AI is helping.

Let's let's by the way, acknowledge when AI failed and let's not make that same mistake again.

So let's not waste time trying to do something with AI that that it can't do.

But things like AI shutouts, things like AI office hours, things like an AI channel in Slack or Teams where you concentrate sort of all that conversation so people can use it as kind of a resource.

All those things work.

They're small, but they're all, they're all part of this change management playbook that we're all going to need to run, you know, to make this happen or, or just be content with this going to happen slower in our going to our organizations and it'll be likely happen faster in others.

This and here's what Toby didn't say in the memo.

It's what he didn't say.

Well, which was frankly the most important thing as an investor in Shopify.

He's going to make the company 20% smaller in, in two years, you know, and I think we need to acknowledge this.

We need to acknowledge this with our employees.

I wish he'd said something.

Maybe not that he can't say that and he doesn't know, none of us really know.

But I think as an investor in Shopify, what I read in that memo was what he didn't say, which is this company is going to keep growing likely and it will be smaller, you know, and or more efficient, more revenue per employee.

That's ultimately for some companies.

I I think a good measure can we drive more revenue with less employees.

And I think in some industries and companies, the answer is obviously yes.

That's another measure.

The last thing I'd say around measuring this is ask employees.

This probably feel about it meaning for our team in section now we're two years in, they, they like having AI around again, it hasn't laid anybody off.

And they, they do like that, like their sentiment is positive around outsourcing the work that they didn't want to do.

So I think it's sometimes, you know, we overplay that like, hey, AI is going to do all the shitty work and and we'll do all the good work.

I don't think that's going to be the case all the time.

And I think, you know, AI would do all the work at some point in certain roles.

But I do think in our in the short term, the next few years, there's a lot to be said for employees who are just a little bit happier in organizations where AI is, you know, helping them in a way that's kind of positive.

This is all really good stuff and, and, and it makes for such an entertaining, dynamic environment.

We're all at the edge of our seats.

I think all the time.

I want to make sure that we can jump into some of the show and tell stuff.

You've been using AI in a really cool way and you were telling me about it before about how you use it as almost like a brain for your for your company.

So I'd love for you to describe what you guys are doing.

And I think that this is something that like everybody should and should try to do for their own company too.

Yeah, for sure.

I mean, we have used AI as a thought partner now for since we started.

Really to me, that was its first and best use case.

And I'm 63.

I type with two fingers.

I want to keep working for 10 more years.

I work and live in Silicon Valley, the most ageist place in the world.

I realized AI could help me keep working and maintain a cognitive edge, and the way it can help me is as a thought partner.

So I've been sort of using I this way from day one.

I think we all should be.

I think it's irresponsible, quite frankly, for any medium to high stakes decision not to be using AI.

And I won't tolerate anyone inside my organization coming to me with a plan or an idea or a request for funds or a request for capital, whatever, without first having talked to AII think it's again, I think it's irresponsible.

And so I used to have a post a note on my screen saying ask AI because I'm not AI native, right?

I'm old.

I grew up in a different world.

You know, the kids in high school and college today, they'll hit the workforce, they're AI native.

I had to remind myself, ask AI use AI.

So I used to use a post it, post it note.

Now what I do is I've got a monitor.

I've got a third monitor on my desktop and that's on a Vertical Horizon.

It's just out of my line of sight, but also in it.

And it's got 1 browser and three tabs open.

Perplexity GPT and cloud, which are my 3 sort of go to AIS.

And so that that was an experiment basically starting in January to see what it would be like to have AAI that present visually that sort of dominant, if you will, on one screen.

And would I use AI more?

And the answer is yes.

And then I also wanted it out of my main screen, my main workspace.

I was trying to understand what it feels like to have AI very present and, and, and lower the friction to work with it, but also maintain my sovereignty kind of over my main, main kind of workspace, right, Which is my Google Docs or, you know, my Asana tasks or whatever.

So that's, that's kind of experiment I'm running.

And I think the answer is from the experiment is we're going to, when AI is easy to use all the time, we're going to use it all the time.

It, it, it will be how we cognitively offload because we all need to cognitively offload.

We all need help, right?

Our brains haven't changed in 70,000 years, but the world around us has changed and the way we work has changed.

It's hard now for our brains to keep up.

So we're going to, we're going to want to offload as much as we can in some respects to AI.

And I think that's generally a good thing.

If you do it the right way, If we have time, we can talk about what is the right way to cognitively offload.

A lot of people are going to do it the wrong way.

They're going to freeload on AI and you can't do that.

You have to manage AI.

You can't freeload.

This is Section Expert.

It was two things I want to show you here today.

This is a clawed project inside of Anthropic's clawed model.

Obviously, a project allows us to upload documents and collaborate with others on work.

So what we thought was let's try this idea of a company's second brain called the Section Expert.

So let's set up a clawed project and we can upload all the documents that we think are relevant to our company this year that other people would want access to and we want the AI to have access to when you're brainstorming with AI and working on sort of a new project.

So what we upload is things like our style guide and our project brief template, our operating principles as a company, our user manuals, how to work with Greg, our H1, OK Rs and so on.

We have our business plan here somewhere and so we wanted really a, you know, a complete picture of section for the AI and we also wanted it to be always the, you know, the right document.

So the chief of staff, our chief of staff maintains this so that the, you know, the right version, the most up to date version of whatever document that is, is inside a section expert.

So we don't have someone, you know, referencing an old plan or something that's now out of date.

So section expert, what we tell the rest of the team is it's up to date, all the right kind of documents and plans and and you know, the pricing model, whatever it might be, is in there.

So when it when you want to brain storm with AI about something related to section about a new project or a new plan, come do it here.

And then you're and you're brainstorming can be public or private.

Obviously your chats can be private or and or public.

So in my case this will be private.

Instead of it, I just want to call out that A, This idea I think is genius, but B, I love how you even put a, you know, working with Greg Manuel into this, into this document.

Yeah.

What kind of things is in the working with Greg Manuel?

Everything from I arrive to most meetings and I arrived to this conversation 5 minutes late.

So I arrive to everything three to seven, 7 minutes late.

Don't take it personally.

My working hours are in the manual in terms of, you know, when you can reach me for what kinds of things, which includes you can text me 24/7 if it's an emergency.

And then here's what an emergency is.

I talk about how I give feedback, which sometimes can be candid and people can, can misinterpret that.

I'm very much oriented to speed.

I'm very much oriented to getting to V1 fast and very much oriented to let's have this conversation efficiently and move on and, and get on to the next decision or the next conversation.

And sometimes that can come across as abrupt or too candid, you know, stuff like that.

Just basically how to how we can faster get to a more productive working relationship.

Same thing for operating principles.

Operating principles, I don't believe in values or, or I think, you know, company values can be so generic and not very helpful in terms of how people really get work done inside the organization.

You know, like, oh, you know, we, we value honesty or truth or, you know, those kind of, I think there's just, they're not helpful.

You know, I, I, I like beliefs and operating principles.

Beliefs are how we think about the world and our place in the world as a company as a start up.

And then operating principles are very concrete that we

typically have between 6:00 and 8:00 depending on the year.

And we usually add one or two and drop 1.

And you know, some are aspirational, I would admit, like we're not there yet, but this is how we want to work.

And some are very much, we work this way.

And so if you, you know, if you come to work here, you should understand these operating principles.

We want, frankly, the AI to understand the operating principles as well, because if they're going to, if the AI is going to brainstorm with us effectively, we want the AI to know how we work.

And so, yeah, I think all these things we want our AIS to know.

We want AIS to have great memory.

I'm excited about open AIS upgrade to memory.

We'll see how useful it is.

But you know, you can see where this is going.

We're seeing the signs of where this is going, and that's one sign to me, which is, you know, bigger memory, much better context awareness and understanding by the AI of our situation.

Obviously the big downside of the AI right now is it doesn't know enough about us.

You know, the context is it's context or situational awareness is limited.

Hence the advice tends to be kind of generic.

So yeah, you can do things to improve that.

You can push the AI when you work with it.

You can add more context, you know, provide more training data and so on.

Really what you want is the AI is just learning about you and your organization all the time.

And so it's context and it's situational awareness will be much higher I think in a year or two when we when we work with it.

So it's going to be great.

I think for AIS being more performant and and more specifically value add to US versus any kind of that generic value.

Certainly, you know, I don't know if you've seen this Gemini Pro feature, it's in their, it's in their development studio.

Yeah, it's not available to consumers, but you basically screen share with Gemini and you know, I haven't got a lot of value from it yet, but you can really see again where it's going.

Yeah, where it's going?

Yeah, the AI is watching, right?

And you are able to interact with it.

So if it was much less friction, in fact, it will likely interrupt you at some point and say, you know, you need to rewrite that or or, you know, that's not a good decision.

I mean again, you can see where this might end up.

So this is really cool.

So it's amazing the whole team have has access to and you were just saying like it's irresponsible for you to not consult AI before you for any major decision.

And so like you, you basically encourage your employees to come here to section expert and then just get a second opinion about anything.

That's right.

Do your brainstorming or thought partnership here.

Yeah, you can do with GPT if you prefer GPT.

But here you know that that Claude has access to everything that that is relevant.

We think in terms of sections business this year kind of thing, or if you're going to use GPT or some other AI you prefer, then load up the business plan maybe or load up the product road map, whatever, whatever it might be.

Don't forget to provide additional context.

And so, you know, these documents are available, you know, to employees.

So, so use them appropriately to get more, get more value from your AI.

But this is going to be the kind of the fastest way to do it if you work at section.

And so this what you referenced is this idea of, of using AI as a board member.

So about 18 months ago when I realized this was really the superpower of AI was as a thought partner, then I thought, well, you know, how do you, how should you think about that?

Well, you should think about as there are decisions you make frequently and AI can be very useful because AI is pretty much free and always available.

It's on your desktop or whatever, it's on your phone.

You can talk to AI.

So there are those decisions, maybe the medium stakes, even maybe the small stakes decisions where you can use AI, you know, as a coach, as a thought partner.

Then there's these high stakes decisions that are likely infrequent, but they they matter a lot.

And that might be a pricing decision in terms of, you know, changing your pricing model as a business, or it might be a hiring decision of an executive hire or maybe even any higher depending on the size of your company.

It certainly is board meetings, you know, board meetings are infrequent, 4 * a year, but they're high stakes for everyone involved.

They're high stakes for the board members, for me, the management team, they take a lot of work and let's optimize our, our, our board meetings was was my point of view and how could AI help.

So what we've did then we've done every every month, every quarter rather since, and we just did it in April before our last board meeting is we ask AI to play the role of a board member and we do that by we give AI the exact same input we give the humans.

So you know, before the board meeting, typically 48 hours, sometimes related than we want to be, you know, we send the board, we send the human board, you know, the board deck is typically probably 40 to 60 slides, something like that.

That's the board's pre read.

And obviously there's an agenda.

We want them to pre read that and then come ready to kind of work through the agenda with us.

Typically 2 hour meeting.

I've got 5 board members.

I've got the best board I've ever had, world class board members in terms of sections board, which is great.

So, but what we do is we basically role play with AI before, before the board meeting.

We do this in part to test the different AI.

So we've been doing this with four or five different AI.

So GPT, Claude, Gemini, Copilot and so on.

And we're trying to get ourselves ready for the board meeting and we're, we're asking AI to adopt A persona.

So a conservative board member, you know, conserving cash, a more aggressive board member trying to drive growth, not worried about cash runway.

You know, a board member that doesn't want to be in the board meeting would rather be golfing.

You know, we've always, we've always got one of those on our board, right?

So we have, we have fun with it and we also really stress test kind of the agenda and what we're going to talk about, you know, prior to that meeting.

So we've also measured the 2.

So since then and every quarter since I've measured the performance of AI to the humans.

Here's how I measure it.

We don't record board meetings.

We don't use AI note note takers and board meetings.

And that's for governance and legal reasons.

That's very common.

So I've got my chief of staff taking, you know, kind of making a great record of what the human said in the board meeting.

We then compare that to what the AIS told us before the board meeting.

And recently, ChatGPT actually went to the top of the leaderboard.

But for the previous year and a half, Claude has been the best performing and performing at 90% of the humans.

So what that means?

Here, can you give us an example?

What kind of things are you asking it?

Just just one example of the sort of thing that you might ask it.

What do you think of the plan?

Is it, is it too aggressive?

Is it too conservative?

How do you think about the cash management?

Is, is this based on the investments we're making is just the appropriate amount of investments we should be making?

What do you think about, what do you think about the sales or go to market plan?

Yeah, anything that's in the.

I mean, this is, this is huge.

Going back to the beginning of our discussion, this has got to be worth more than $20.00 a month.

I'm just saying.

Yeah, absolutely.

And really I don't even I can ask specific questions and we do do that.

But really what you can do to get started, to start the conversation is just say listen, pretend to be a board member.

So you ask the AI to adopt A persona, you know, as a, as a board member.

That's enough to get started.

And then you can go deeper in that conversation with, OK, let's let's pick parts of the plan or, you know, a specific topic to discuss.

But then initially just start the conversation with I'm the CEO, you're the board member and I want you to play a conservative board member, you know, really focused on preserving value and minimizing risk.

You know you're going to have a good conversation just with that prompt persona adoption.

Buy the AI.

It's just an important little hack or pro tip for all of us using AI as a thought partner.

You know, when you're, when you're asking for medical advice, tell the AI you're an oncologist at the Mayo Clinic or you know you're an orthopedic knee surgeon at Vail.

You know, whatever it might be, like give the AIA persona and push the AI to really, you know, adopt that persona and give you the kind of advice you want.

So yeah, we, we, we do this every board meeting.

So the, the AI performs 90% at the human level.

What's that mean?

It means they, the AI gives us 90% of what the board's going to tell us.

So, you know, and then the 10%, let's talk about that in the 10%.

And we analyzed the gap where, where the AI and the board didn't say the same things.

The AI is telling the stuff the board doesn't say, which is just interesting to know that the AI will come up with unique insights that the, you know, the humans didn't and vice versa.

The humans are providing value.

As you can imagine, it's human value.

So what humans are great at are networks and connections.

And by the way, they write checks, right?

If they're on your board, if they're investors yet, not yet, but soon.

AIS don't write checks.

So you know, you need your humans on site if they're a source of funds, but they'll be able to have more context in the moment.

So at this moment with economic chaos emanating out of Washington based on the tariffs, you know, board members are great, AI is not so good in the moment.

And instead of relating to what's happening in the world around us, what's happening to the funding environment, what's happening to, you know, the customer environment, which might be the buyer environment if you're, if you're AB to B company.

So boards have that 10%.

That's where I want them to focus.

And by the way, when I send the board deck to the humans, I give them their prompt to use AI to do the same thing on the other side.

Like don't come to a board meeting without having talked to your AI about this company.

As a minimum, read the board deck.

That's a minimum, right?

Try to be prepared.

Read the pre read.

My point of view is every board member and I make it easier for them because I send them the Proms.

They should be doing what I'm doing.

They should be working with their AI and say listen, pretend to be the CEO.

I love this.

This is so cool.

You know, it kind of reminds me that like the parallels here are, you know, we talk about, you know, GitHub copilot or Devon or some of these tools that will comment on your code.

So you submit APR and you can get a comment from like these other.

This is like a comment on on your business, right, You have new plan.

Here's like, you know, all the quarters results and and they can just comment on it and give you things that you didn't think about.

This is, I mean, this is really cool and there's all these people that for example, you.

So that might listen and they'll say like, hey, I don't have a board.

You know, we're not a venture backed business.

We're just like owner operated business.

And so, but now you can create an AI board and I think that's just incredible.

Absolutely.

I think that's, that's I think almost the best use case actually, which is I don't have a board, but I want, I want AI to play this role of a board member.

And again, a specific kind of board member.

If you, if you are all about in terms of your own business, preserving value, preserving, you know, generating cash, then play that kind of role of as a board member, someone who's focused on, you know, generating cash versus generating growth, for example.

Or if you're frustrated with your own sort of growth of your, of your own company and you, and you have a hard time kind of coming up with growth ideas or sort of pushing into the risk associated with growth, then ask your AI to play that kind of board member or that kind of coach for you.

Like what are some growth ideas and what are the riskier ones and so on.

And could you help me sort of think through how to make one of those growth oriented moves because some of us are just naturally more conservative.

The good news about AIAI doesn't have blind spots.

We have blind spots.

Yeah, this is great.

I mean, this is this is amazing.

I'm I'm conscious that we're we're running against time.

So yeah, I would love to see an example of what what this types out as this is, this is going there.

You know, the the final question that we always like to end on.

I mean, this is incredible as I see this like going through and assessing your business plan.

Yeah, I mean, look at this right again, I probably knew some of this or it's all floating around in my head.

I mean, I'm an experienced CEO, 7 time, 7 time founder and CEO so but even, and then it's great to see it sort of written out.

It's great to see, see it summarized.

Oh yeah, there are all these risks, right?

We again have blind spots.

Entrepreneurs tend to be more optimistic than pessimistic, right?

We tend to think about one risk.

Well, have you thought about the other two risks?

Again, it's really helpful to have a coach, you know, a mentor, a board member, whatever you want to call it, just with no blind spots, no emotional charge.

I also think we forget about when we get advice from humans, it comes with an emotional package.

It comes with an emotional wrapper that's hard for us.

It could be good, it could be bad, it doesn't matter.

My point is it's coming from someone.

So it it arrives in on our side, you know, not just the advice, but the advice inside of that emotional wrapper.

Who's this advice coming from?

Someone we admire, someone we trust or don't or and so on.

We don't have that with AIA is got this purity to it.

And I think as a coach or mentor that, you know, the advice doesn't come with, it's not loaded with anything.

You know, it's just what it is, right?

This is it.

This is incredible.

Thank you so much for showing it to me and the audience here.

So final question for you, what are you most excited about over the next 12 months?

I'm most excited about the fact that because I'm seeing it in my company and I'm starting to see it elsewhere, that people who don't necessarily look like they can take that job, whatever that job is, can take that job, meaning, you know, get the job or do the job because they're using AI.

I'm very optimistic about how people would, without the right educational background, or maybe they're not a software engineer or a data analyst, but now they can get their own data and do their own data analysis.

Whereas before they were kind of waiting and depend upon someone else.

And then they can kind of flex and grow into kind of a more valuable position or take on more work in a good way.

It's the optimistic side of this.

Again, there's the other side of this coin, but I'm most excited about how this really does enable people, if they can use these tools the right way, to step into roles and opportunities and tasks and get done that they couldn't do before and I think be viewed as more valuable.

So if if someone wants to take a course on section or find you on the Internet, where would you send them?

Sectionai.com So like a lot of other companies, I've added AI to the and yes, that's because I want an AI valuation.

So yeah, we're section ai.com or e-mail me greg@sectionai.com And listen, we we can provide a discount code for your listeners, so we can put that in the show notes as well.

So you'll find the discount code in the show notes.

Greg, thank you so much for coming on the show.

Yeah.

Thank you.

It was great.

And that's it for today.

Thank you so much for tuning into this episode of this new way.

If you like the content, be sure to rate, review and subscribe so you can get notified when we post the next episode.

See you next time.