Caleb.AI Scorecard
← Caleb’s latest thinking
AI and business · Leadership · The company brain

What Is a Company Brain, and Why Does Your Business Need One?

By Caleb Storkey, AI consultant · 10 min read

The AI models are ready. Now your business knowledge needs to catch up.

In short

A company brain is a living layer of your business's knowledge, decisions and ways of working, kept current and structured so that both people and AI Employees can act on it safely. In 2026 it has become the missing piece between buying AI tools and getting real results from them. Building one is mainly a leadership job: deciding who owns it and keeping it true.

01The blocker has moved

Every business I speak to is somewhere on the same road with AI. Some have barely started. Others have a dozen tools and a nagging feeling that none of them really understand the business. For three years most of the attention has gone on the models, and they’ve improved at a startling pace. For most businesses, the model is no longer the thing holding them back.

Many people compare AI to electricity. If that is right, the question is no longer whether AI matters. The question is what you build with it.

Listen to the people closest to AI this year and you hear the same message. Microsoft's 2026 Work Trend Index found that most AI users are ready, but the organisations around them have not redesigned how the work gets done. Y Combinator, the best-known startup accelerator, put it more bluntly: the biggest blocker to automating companies with AI is no longer the models. It is the domain knowledge.

That knowledge is everything your business knows that has never been written down in one place. It sits in inboxes, Slack threads, shared drives and call recordings. A lot of it sits in people's heads. If one of your key people handed in their notice on Friday, a surprising amount of how your business works would walk out of the door with them.

People cope with that mess because they half remember where things are and who to ask. An AI Employee cannot do that.

An AI Employee works with exactly what it is given, and nothing more.

The research houses now say what founders feel in their gut. In February 2026, Gartner reported that a lack of data readiness is one of the top barriers to AI, and that more than three-quarters of organisations are now prioritising investment in AI-ready data. A March 2026 study by Cloudera and Harvard Business Review Analytic Services went further.

Evidence · data readiness
7%
of enterprises say their data is completely ready for AI.Source: Cloudera and HBR Analytic Services
Cloudera and Harvard Business Review Analytic Services, “Taming the Complexity of AI Data Readiness”, March 2026: a survey of more than 230 people involved in their organisation’s AI data decisions. Read the source →

I love working with the early movers in that 7%. That is where a lot of the fun and the new ideas happen. But most of my work is with the other 93%.

02The idea is older than AI

The vision behind the company brain is more than fifty years old. Only now do we have the tools.

In 1972, a British management thinker called Stafford Beer published Brain of the Firm. He modelled a business on the human nervous system: senses that notice what is happening, reflexes that respond, and a brain that learns and steers. It was one of the first serious attempts to describe a business as something that thinks.

He also tried to build one. In 1971, Chile's government brought Beer in to create Project Cybersyn, a network of telex machines linking factories to an operations room in Santiago. In October 1972, when around 40,000 striking truck drivers blocked the roads, the system helped keep essential goods moving. It was a company brain made of telex paper and slide projectors, decades before anyone had the tools to do it properly.

How it works · fifty years of the company brain
  1. 1972Stafford Beer publishes Brain of the Firm
  2. 1972Project Cybersyn links Chile’s factories
  3. 1988Drucker: the business as an orchestra
  4. 2025“Context engineering” replaces prompt engineering
  5. 2026Gartner calls context “the brain for AI”
The idea came first. The tools arrived in the last two years.

Sixteen years later, Peter Drucker picked up the thread. In his 1988 Harvard Business Review article, "The Coming of the New Organization", he predicted that the large business of the future would look less like a factory and more like a hospital or a symphony orchestra. It would be organised around information. In an orchestra, he pointed out, every specialist plays from one shared score, with no layer of middle managers in between.

Hold on to that picture. We’ll come back to it.

What Beer and Drucker lacked was a way to capture what a business knows and put it to work at the speed of the business. Now we have one. In June 2025, Shopify's Tobi Lütke said he preferred "context engineering" to "prompt engineering", which he described as giving a model all the context it needs to plausibly solve the task. Andrej Karpathy backed him within days. What matters most is what the AI knows at the moment you ask it.

I really like the term “context engineering” over prompt engineering. It describes the core skill better: the art of providing all the context for the task to be plausibly solvable by the LLM.
Tobi LütkeFounder and CEO, Shopify · on X, 19 June 2025

By 2026, the analysts had arrived at the same place. In April 2026, Gartner framed context as critical infrastructure for enterprise AI and described it as "the brain for AI". Vendors are racing to sell it and investors are funding it. Garry Tan, who runs Y Combinator, has even open-sourced the brain behind his own AI agents.

Twenty years of startups have taught me this. When an old idea finally gets its tools, it becomes standard kit almost overnight.

03Solar brain or company brain

Here is how I explain it when I am in the room with a leadership team.

Most companies run on a solar brain. Every department is its own sun. Sales keeps its truth in the CRM. Finance keeps its truth in the accounts. Operations keeps its truth in someone's notebook. Each one hoards its light and fights for budget.

Then AI arrives, and most businesses give every department its own chatbot. Now there are seven suns with seven AI assistants, each one answering confidently from half the picture.

A company brain flips that. There is one shared intelligence, and every team and every AI Employee orbits the same truth. It keeps improving, because every call, decision and lesson goes back in and makes the next piece of work better.

How it works · one shared brain
Every team and every AI Employee works from the same truth, and what they learn flows back in.
A document library starts going out of date the day after it is written, while a brain keeps learning.

04The Four Rooms and the Librarian

When I walk teams through this, I use a simple picture. A company brain has four rooms and a librarian.

How it works · the four rooms and the librarian

Room one is the Library. It holds what you know: your offers, prices, customers, products and policies. Most businesses have a Library of sorts, but it is scattered across ten tools and three versions.

Room two is the Ledger. It holds what you decided, and why. This is the room almost nobody builds. In December 2025, the investors at Foundation Capital pointed out that today's systems each hold one piece of a project, and none of them capture why decisions were made. The why is what stops an AI repeating last year's mistake with total confidence.

Room three is the Playbook. It holds how you do things here: your voice, your standards, what "good" looks like, and the steps behind a proposal, an onboarding or a complaint. This is where knowledge becomes something an AI Employee can carry out.

Room four is the Pulse. It holds what is changing right now: the new client, the new price, the supplier who let you down last Tuesday. Without the Pulse, the brain slowly falls out of date.

Then there is the Librarian. Garry Tan describes a company brain as the library plus the librarian, the part that decides which three books get opened for any given task. I would add a human librarian too: a named person who owns what goes in, what comes out, and who is allowed to see what. That person doesn’t have to be you.

Most businesses have a messy Library, no Ledger, a Playbook that lives in one person's head, and a Pulse that arrives by email at eleven at night.

05The role of the AI Conductor

The librarian keeps the brain true. Someone still has to lead with it, and this is where the conductor comes back in.

Picture a hundred musicians, each brilliant on their own instrument. Without a conductor and a shared score, you get noise. With them, you get something none of them could play alone.

A business with a company brain works the same way. The brain is the score. Your people and your AI Employees are the musicians, each one a specialist. And the leader becomes the conductor.

Real orchestras have a librarian too, who makes sure every player has the right part with the conductor’s markings in it. In a business, that’s your human librarian. They keep the score current, so the conductor can give their attention to the music.

The conductor does not play every instrument. They couldn’t, and they shouldn’t try. Their job is different:

The conductor’s four jobs
Set the tempoThey decide what matters now and what waits.
Pick any job to read it on its own.

Benjamin Zander, the conductor of the Boston Philharmonic, likes to point out that the conductor does not make a sound. Their power comes from making other people powerful.

That’s the shift a company brain asks of a leader. In the old model, the leader sat at the top of a chain of managers who passed information up and down, losing a little at every step. In the new model, the leader stands in front of the whole orchestra, everyone reads from one shared score, and every player can see the conductor.

In practice, the AI Conductor spends less time firefighting and more time listening, which is a far better use of a leader's gifts.

There has been a lot of talk about Founder Mode since Paul Graham's 2024 essay, inspired by Brian Chesky at Airbnb, about founders staying close to the detail. I think of what comes next as Conductor Mode. You stay close to the music without playing every part. You know which instruments need your attention right now, and you listen for how the whole piece blends together.

06Why most attempts stall

This is the hard part.

Building a company brain sounds like a weekend job with the right tool. It isn’t.

You could fairly say this is knowledge management with a new name. Businesses have built intranets and wikis for thirty years, and most of them ended up as digital attics. Two things are different now. The main reader is no longer a busy person skimming a page; it is an AI Employee that acts on what it reads, many times a day. And the brain can be kept current as the work happens, instead of relying on someone remembering to update a document.

Most attempts stall on the upkeep.

These days, the technology is the easy part.

The hard parts are human and strategic:

Six places it stalls
Deciding what goes inIf you pour in everything, the AI drowns in noise and old versions. If you put in too little, it guesses.
Pick any point to read it on its own.

Microsoft's 2026 Work Trend Index points the same way. 58% of AI users say they are now producing work they could not have done a year ago. Yet only 26% say their leadership is clearly aligned on AI, and only 13% feel rewarded for redesigning how work gets done.

Evidence · ready people, unready systems
58%of AI users say they are producing work they could not have done a year ago.
26%say their leadership is clearly and consistently aligned on AI.
13%say they are rewarded for reinventing work with AI, even when results fall short.
Source: Microsoft Work Trend Index, May 2026
Microsoft, 2026 Work Trend Index Annual Report, 5 May 2026: a survey of 20,000 AI users in 10 countries, fielded by Edelman Data x Intelligence. Read the source →

The analysts see the same gap from the other side. Forrester's State of Agentic AI report, published in June 2026, found that about three-quarters of enterprise leaders say they are adopting agentic AI, yet only a small minority run it in real, everyday production. Gartner's 2026 CIO survey puts the share of organisations that have deployed AI agents at just 17%.

The businesses that do succeed spend their money differently. In April 2026, Gartner reported that organisations with successful AI initiatives invest up to four times more, as a share of revenue, in foundations such as data quality, governance and change management. In other words, they put money into the brain as well as the agents.

Evidence · what successful AI spends on
Up to 4x
more invested in data quality, governance and change management by organisations whose AI succeeds.Source: Gartner, April 2026
Gartner press release, 16 April 2026: a global survey of 353 data, analytics and AI leaders. Read the source →

The people are ready. The structure around them isn’t.

07What changes when it works

When the brain is in place, three things shift.

People get up to speed much faster. That goes for new staff and for new AI Employees. What the business knows is no longer trapped in one head, so nobody has to wait for that person to come back from holiday.

AI stops sounding generic. It quotes your prices, writes in your voice, follows your rules, and knows why you stopped working with that supplier. The output starts to sound as if it came from inside your business.

The leader gets their head back. You stop being the human search engine. You go fishing in the river when you need something, instead of being the river.

Gartner's April 2026 research backs this up. It found that organisations with the most mature AI-ready data achieve up to 65% greater business outcomes, including revenue growth and lower costs.

Evidence · what AI-ready data delivers
Up to 65%
greater business outcomes, including revenue growth and lower costs, for organisations with the most mature AI-ready data.Source: Gartner, April 2026
Gartner press release, 16 April 2026: a global survey of 353 data, analytics and AI leaders. Read the source →

In my own business, 40 AI Employees run alongside me. Every one of them is only as good as what it knows about the business, and only as safe as the person who keeps that knowledge true.

The company brain is a leadership decision first.

It’s the moment a business writes down who it is and how it works. That’s where the threads start coming together.

08Put it into practice

You could build this alone. I wouldn’t.

The businesses that stall treat the company brain as a tool to install. The businesses that win treat it as a redesign of how they work, with someone alongside them who has done it before and knows where the drop-off points are.

Helping business owners and managers become AI Conductors is what my business is built around.

If you want help building the rest, come and find me. The company brain is one of the modules in my accelerator, and it is also where our audits start.

For bigger builds, my team builds it with you. We bring the tech, and I bring what it takes to make the change stick. We will map your four rooms with you, decide what goes in first, and stay to make sure it keeps working.

Whatever size your orchestra, we’ll help you conduct it.

Before you do anything else, take a breath and run the PIE check.

A slice of PIE to help with building your company brain

The PIE check · a slice at a time
  1. P · Personal.Which knowledge in your business lives only in your head?
  2. I · Internal.If your best person left, which room would empty first?
  3. E · External.Which question do customers keep asking that your team answers differently every time?

If those answers are fuzzy, start with one room this week, and give it an owner and a review date.

Key takeaways
01
The model is no longer the blocker.What holds most businesses back is knowledge that has never been written down in one place.
Free live webinar

Cut Your Workload from Your Week, Every Week

Hand off the admin. Get your evenings back.

Monday 23 November 15:30 to 16:30 GMT (UK) · 10:30am to 11:30am ET (US) Live on Zoom
Find Out More

* Required.

Source notes

Microsoft, 2026 Work Trend Index Annual Report, “Agents, human agency, and the opportunity for every organization”, 5 May 2026 (20,000 AI users in 10 countries). Source of the 58%, 26% and 13% figures. Link

Y Combinator, Requests for Startups, Summer 2026, “Company Brain” (Tom Blomfield). Link

Gartner, “2026 CDAO Priority: Evolve the AI-Ready Data and Analytics Architecture”, 17 February 2026. Link

Cloudera and Harvard Business Review Analytic Services, “Only 7% of Enterprises Say Their Data Is Completely Ready for AI”, press release, 5 March 2026. Link

Stafford Beer, Brain of the Firm, 1972. Link

Project Cybersyn, Chile, 1971 to 1973. Link

Peter Drucker, “The Coming of the New Organization”, Harvard Business Review, January to February 1988. Link

Tobi Lütke, post on X, 19 June 2025. Quoted word for word. Link

Andrej Karpathy, post on X, 25 June 2025. Link

Gartner, “Gartner Says Organizations with Successful AI Initiatives Invest Up to Four Times More in Data and Analytics Foundations”, press release, 16 April 2026. Source of “the brain for AI”, the up to 4x figure and the up to 65% figure. Link

Garry Tan, GBrain, open-source repository on GitHub, April 2026. Link

Jaya Gupta and Ashu Garg, Foundation Capital, “AI’s trillion-dollar opportunity: Context graphs”, 22 December 2025. Link

Garry Tan, “Every company should have a Brain”, talk at AI Engineer World’s Fair, published by Y Combinator, 17 July 2026. Link

Benjamin Zander, “The transformative power of classical music”, TED, 2008. Paraphrased, not quoted. Link

Paul Graham, “Founder Mode”, September 2024. Link

Forrester, “The State of Agentic AI in 2026: Companies Are Chasing, Few Are Catching”, 3 June 2026. Link

Gartner, Hype Cycle for Agentic AI, 15 April 2026, citing the 2026 Gartner CIO and Technology Executive Survey. Link

Written by Caleb, co-author of Futureproof.

Questions people ask

What is a company brain?

A company brain is a living layer of your business’s knowledge, decisions and ways of working, kept current so people and AI Employees can act on it safely. It is more than a document library because it keeps learning from every call, decision and lesson.

Is a company brain just knowledge management with a new name?

Partly. The difference is that its main reader is now an AI Employee that acts on what it reads many times a day. The brain can also be kept current as the work happens, instead of relying on someone remembering to update a document.

What should go into a company brain first?

Start with one of the four rooms. Most businesses begin with the Library (offers, prices, customers and policies) or the Playbook (how things are done). Put one person in charge of keeping it current.

Who should own the company brain?

A named person, not a tool. They decide what goes in, settle conflicts between versions, and control who can see what.

Why do AI tools give generic answers about my business?

They only know what they are given. Without your prices, voice, rules and past decisions, they guess. A company brain gives them that context.

How can an AI consultant help build a company brain?

An AI consultant such as Caleb Storkey maps your four rooms with you, decides what goes in first, and sets up permissions and upkeep, so the brain keeps working after launch.

Take it further

Your learning log · 0 of 8 points read

Related reading

Read Caleb's Latest Thinking →