AI for professional services
We speak the languageAgencies, consultancies, and expert firms where the product is judgment and the constraint is the founder's calendar.
In a professional services firm the product is what the senior people know, which means the ceiling is their calendar. The fix is not automating the thinking. It is capturing the knowledge that only lives in their heads, so the rest of the firm can work from it and the deliverables stop being rebuilt from scratch.
Our own
This firm runs on what it sells
7 notes
The company brain, in plain markdown
You own it
Your knowledge, your cloud, always
Where the time goes
We know this one from the inside, because we run a professional services firm and every one of these has bitten us.
The senior person knows how the firm prices, what to do when a client goes quiet, and which vendor to avoid. None of it is written anywhere. Every junior question routes to them, so the firm's throughput is bounded by one person's availability, and their calendar is the actual product.
The firm's ceiling is one calendar
You have made this document forty times. Every time somebody opens a blank page, finds the last decent version, copies it, and forgets which parts were specific to that client. The intellectual work took an hour. The assembly took four.
The assembly costs more than the thinking
A qualified lead is waiting while somebody assembles scope, pricing, timeline, and the relevant case study. Most of it is decisions the firm already made on ten similar engagements. It takes a week because nothing holds those decisions except memory and a folder of old proposals.
A week of waiting on assembly work
An hour on the phone with a client produces real insight about their operation, their vocabulary, and their fear. It ends up in a note somebody writes at 6pm, or nowhere. Next quarter, someone else asks the same discovery questions to the same client.
The insight ends at 6pm or nowhere
A new hire is billable when they have absorbed how the firm actually thinks, and that transfer happens by osmosis: sitting in on calls, getting work returned, and slowly learning the rules nobody wrote. Six months of somebody's salary is the price of not having written it down.
Six months to absorb the unwritten rules
Somebody tried a model on the firm's work and it produced generic output, because it did not know how your firm thinks. So the conclusion was that AI is not there yet. The real conclusion was that a model with no knowledge of your firm produces the average of the internet, which is exactly what it did.
Generic in, generic out
The machine
The machine underneath an expert firm. Notice how much of it is knowledge moving between people rather than work getting done.
01
02
03
04
The split
In this industry the temptation is to automate the thinking, because the thinking is what takes the time. That is backwards. The thinking is what clients pay for, and a firm that automates it becomes a cheaper version of itself. Automate the assembly, the recall, and the transfer. Keep the judgment expensive and give it more room by taking the four hours of formatting off it.
What we build
01
Seven plain markdown notes, in your own cloud, holding how your firm operates: the method, the pricing logic, the vocabulary, the constraints, the decisions and why. Everything else on this list reads from it. Without it, every AI system you build produces the average of the internet with your logo on top.
02
A discovery call becomes a structured record: the client's language, their constraints, the bottleneck, and the thing they said that changes the engagement. Not a transcript. A transcript is a recording of a call that still needs somebody to read it.
03
The document drafted from your method and this client's context, so the expert edits rather than assembles. The hour of thinking stays yours. The four hours of putting it in the right shape is the part worth taking back.
04
Scope, pricing, timeline, and the relevant proof, drafted from what comparable engagements actually took. A week becomes an afternoon, and the number is anchored on your history instead of your optimism.
05
Juniors ask the brain first. It knows how the firm prices, what happens when a client goes quiet, and which decisions were already made and why. The senior person answers the questions the brain cannot, which is a much shorter list than their current inbox.
06
One knowledge base becomes the ten working documents different people actually need: the onboarding doc, the pricing sheet, the client-specific brief, the process note. Written once, kept current in one place, rendered wherever it is needed.
In production
CMore Flo · Professional services
Our own product / live and in daily useEvery AI system we build reads from a company brain: a small set of plain markdown notes, in the client's own cloud, holding how their business actually works. We run on ours. It is live at brain.cmoreflo.com, the starter is free, and this page is the honest version of what that is.
7
Plain markdown notes, no proprietary format
Your cloud
We do not host your knowledge
Free
The starter, with no call required
Under the hood
Knowledge and docs
Client and pipeline
Calls
The company brain goes in whichever of these your firm already lives in. It is plain markdown on purpose, so it is readable by a person, portable to any tool, and not dependent on us or on any particular model. The point is that it stays yours when everything else changes.
Principles
This firm operates on the same company brain, the same human checkpoints, and the same eval gates we build for clients. That is the honest version of proof for professional services: not a case study about somebody else, but an operation you can ask hard questions about. We publish the method for free on the how we work page for the same reason.
We give away the method, the stack, and the prompts, because none of that is the advantage. The advantage is the loop: the knowledge that compounds, the tests that keep it honest, and the operator who curates it. A competitor can read everything we publish and still not have your firm's brain.
A knowledge base that accepts whatever the model decided was worth remembering becomes a landfill in a quarter. There is a gate: a person approves what becomes true. It is the least glamorous part of the build and it is the part that decides whether any of it works in month six.
The thinking is the product. The formatting, the recall, and the rebuild-from-scratch are not. A firm that automates the first thing has made itself cheaper. A firm that automates the second has made its experts faster at the thing clients actually pay for.
Same approach, different language
Commercial real estate
Commercial brokers working multifamily, retail, and hospitality, whose day still runs on a cold-call list.
We speak the language →Legal
Small and mid-sized firms where the partners are the product and the admin load is uncapped.
Patterns, no client yet →Financial services
Advisory firms, lenders, and finance teams where senior people spend their day gathering what they need to think about.
Patterns, no client yet →Questions
It is a small set of plain markdown notes, in your own cloud, that hold how your company actually operates: your method, your pricing logic, your vocabulary, your constraints, and the decisions you have already made with the reasons attached. Every AI system you build reads from it. It is the difference between a model that sounds like your firm and one that sounds like the internet. It is also the first paid deliverable in most of our engagements, and there is a free starter version if you want to build one yourself.
Not the part clients pay for, and if you aim it there you will make your firm worth less. The thinking is the product. What it should replace is the four hours of assembly around the one hour of thinking: rebuilding the deliverable, recalling what a comparable engagement cost, and answering the junior question that a written method would have answered.
Because the model did not know anything about your firm, so it produced the average of everything it had read, which is by definition generic. That is not a limitation of the model, it is a missing input. A model with your method, your pricing logic, and your client's context in front of it produces something recognizably yours. Almost every abandoned AI experiment we hear about is this, misdiagnosed as the technology not being ready.
Yes, and it is the only proof we can honestly offer in this category. This consultancy runs on the same company brain, the same human checkpoints before anything irreversible, and the same eval gates we build for clients. We do not have a third-party professional services case study, and we are not going to invent one. The construction and insurance pages are where the shipped client systems live.
Somebody curates it, and that is a real job rather than a background process. Your firm changes: a price moves, a method gets revised, somebody makes a decision in a hallway. If that never lands in the brain, every system reading from it slowly starts lying. The consolidation gate is what keeps it short and current instead of letting it grow into an archive nobody trusts.
It stays in your cloud, in your accounts, under your access controls. That is the architecture rather than a policy: we do not host your knowledge base, so there is no version of this where your client work sits on our infrastructure. For anything sensitive, the rule we run internally is that it stays local and it stays out of any repository.
The company brain is usually the first module and it takes days rather than months, because the raw material already exists in your heads and your folders. It is useful the moment a junior can answer their own question from it. Everything else on this page is downstream of that, and none of it works without it.
Yes, and we would argue the opposite of what you are worried about. We publish our own method for free. The advantage was never the secret, it was the loop: the knowledge that compounds, the tests that keep it honest, and the person who curates it. A method nobody in your firm can read is not protected. It is just stuck in one person's head, where it cannot scale and cannot outlast them.
Read next
Explainer
What an AI-legible business actually looks like, the five-point readiness checklist to score your own operation, why one person is usually compensating for every broken handoff, and the 90-day path from messy to legible.
Explainer
What does an AI consultant actually do? A plain-English breakdown of the real work behind operational AI consulting, from diagnosis to system design to ongoing results.
Strategy
Why business operators should build portable markdown knowledge bases before relying too heavily on any one AI model, platform, or vendor memory layer.
Where to start
Find the bottleneck, price the fix, build the system, then keep compounding it. The four engagements are how you buy it, and you can start on any rung.
See the four engagements →Get started
Book a 30 minute call. We find the one place your work waits on you, and you get a straight answer about whether AI is the fix. No pitch, no deck.