Get 25 battle-tested prompts
Turn real client conversations into Linkedin content that builds authority and attracts the right clients.
Turn real client conversations into Linkedin content that builds authority and attracts the right clients.
A few weeks ago I was on a call with the founder of a GTM consultancy. He had just finished building his firm's AI brain and was about to put it in front of his first client. Smart guy, has sold consulting for years, closes deals in his sleep. He stopped mid-sentence and asked me: "so how are you thinking about positioning this?"
He wasn't asking about Gia. He was asking about his own product. He had something that works for his clients at 10pm on a Tuesday, something that knows his frameworks and their business better than any junior on his team, and he had no idea what to put on the invoice.
That's the problem. If you run a consulting firm, an agency, or an implementation shop, you have sold exactly one thing your whole career: time. Hours, retainers, project fees that are hours in disguise. Every pricing conversation you've ever had was anchored to how much of you the client gets. Now you've got a thing that gives them all of you, all the time, and none of your instincts apply. It feels like selling software, and you've never sold software.
I've had somewhere north of 900 conversations with professional service firms in the last couple of years, and since we pivoted Gia to help firms turn their expertise into a client-facing AI product, this exact question comes up on nearly every call. Not "will my clients want it" (they do) and not "will it work" (it does). The question is: how do I charge for this, how do I talk about it, and what do I call it?
This post is my attempt at a straight answer. It's built from the firms we work with, what they've tried, what stuck, and where it's still unsolved. It's a playbook, not a sales pitch. If you're building this on your own stack or with someone else, most of it still applies.
Before the how, the why. Because the first time you try to price this you will be tempted to give up and just throw it in for free. Sometimes that's right (more on that below). But you should know what you're walking away from.
Revenue that isn't tied to hours. This is the obvious one and it's still the big one. Every dollar you've ever made required someone on your team to be in a room, on a call, or in a doc. A client-facing brain earns while nobody is working. You can only take on as many clients as you have hours, and the only way to scale has been to hire. People have been writing productization playbooks for consultants for a decade; this is the first version where the product actually does the work.
Margin. When a client asks a question between meetings, someone on your team answers it. That's unbilled time, or it's billed time that annoys the client. One firm we talked to had the classic version: after a project wraps, the client keeps asking questions, and the firm either answers for free or says "you'll need to buy five hours a month of support." Neither option is good. A brain answers those questions at near zero marginal cost. The margin on the engagement goes up without touching the price.
The relationship doesn't end when the project does. A friend of mine who runs a consultancy described it as a "leave behind." The deliverable used to stay and the reasoning left. Eighteen months later someone asks why the system was set up this way and the people who knew have moved on. If your brain stays with the client after go-live, you're still in the building. The next project comes to you.
You see what clients are thinking. This one surprised me. Several firms told me the observability is worth more than the revenue, at least early on. You see every question a client asks at 10pm. What they're confused about, what they're planning, what they're about to ask you for. One founder called it "getting your eyes on what they're struggling with" between calls. That's upsell intelligence you've never had.
Every consulting firm is getting compressed. An advisor we work with put it plainly on a call: every firm is thinking about a new revenue stream because they're getting squeezed in some shape or form. Clients have ChatGPT. They're asking harder questions about what the retainer is for. Even the Big Four are rethinking the billable hour because AI is doing in minutes what they used to bill by the day. A product that carries your expertise is a real answer to that, and it's one most firms haven't figured out yet. Being early matters here.
The single most common mistake I see is firms describing the thing by its technology. "We built an AI assistant." "We've got a chatbot trained on our docs." "It's an LLM with RAG over our knowledge base." Your client hears "software" and immediately compares it to ChatGPT, which they already pay $20 a month for.
Don't open with the technology. Lead with what it knows.
Here's the one-sentence version we give firms, and it's held up well in real pitches:
"Everything we know about your business, and everything our firm has learned over the years, in something your team can use directly at any hour."
If you've got thirty seconds instead of one sentence:
"We've built a brain for your project. It's read the plan, the docs, our calls, every decision we've made and the reasoning behind it. It sits on top of everything our firm has worked out across every engagement we've run. Your team can ask it anything they'd normally send us and get an answer straight back, with a link to whatever it came from. It lives in your Slack, so there's nothing new to log into."
Notice what's doing the work there. Not "AI." Not "trained." It's "every decision we've made and the reasoning behind it" and "everything our firm has worked out across every engagement." That's the thing ChatGPT cannot do, and your client knows it instantly.
Name it after your firm, not after the tech. Every firm we work with white-labels it. A GTM consultancy calls theirs "[Firm Name] AI." Another firm we work with does the same with their name on it. The client doesn't see Gia and shouldn't. When a client asks "did you build this?" the honest answer is: we built everything in it, our methodology, our standards, everything from your project. The platform underneath is a specialist provider we work with, same as we use HubSpot or Asana. Nobody asks a consulting firm whether they wrote their own CRM.
Where it goes in a new business pitch. Not near the front. Nobody picks a firm because of a tool, and leading with it makes you look like you're selling software. Bring it in when you're talking about how you work with clients, or what happens after go-live. Roughly two thirds of the way through. The line that works when you're up against similar firms:
"Most of what you're hearing this week will look fairly similar. Same methodology, same kind of team, same shape of plan. The difference here is that you're not just getting the four people we put on your account, you're getting everything this firm has learned across every project we've run. And your team can get at it directly."
The ChatGPT question. You will get asked "isn't this just ChatGPT?" Every time. The answer is yes, same underlying models, and the difference is what it has access to and what it's allowed to do. ChatGPT will only ever give generic advice. This has read your project. It knows why we chose option B in March. It won't make up an answer and it won't quote you a price. And frankly it's more secure than your team pasting your data into a consumer subscription.
One more thing on positioning that took me a while to see. The pitch is not "we're giving you AI." The pitch is "we're giving you more of us." Clients hire your firm for judgment. This is your judgment, available when you're not in the room. Frame it that way and the price conversation gets a lot easier.
Here is every way I've seen a firm package a client-facing brain so far. Not theoretical. These are the ones that showed up across our customer calls. Most firms end up using two or three of them at once for different clients.
The simplest version. The brain becomes part of the engagement. It's not a line item, it's just how you work now, and the project fee goes up to reflect it. The pitch is that the client gets 24/7 access to the firm's thinking for the life of the project, so the engagement is worth more.
This is the easiest to sell because there's nothing new to buy. The client is already saying yes to you. It works best for new engagements where you're setting the price from scratch. It's awkward on existing accounts, because raising a retainer mid-stream needs a reason, and "we built a thing" isn't always enough of one.
Same engagement, but the brain shows up on the proposal as its own line. "Project brain access: $X/month." A system integrator we work with does exactly this. They build the AI brain as part of the implementation and the client pays a monthly fee for it, listed separately.
The upside is clarity. The client knows what they're paying for and you can see whether they value it. The downside is you've invited a negotiation. A line item can be crossed out. If you're going this route, make the number small enough that nobody fights it and large enough that it's real. A few hundred a month is where most firms land.
This is the one I find most interesting, and it's where the recurring revenue lives.
The project wraps. Normally the client goes dark, or they stay on a full retainer they don't really need. Instead you offer a lower monthly fee to keep access to the project brain plus your firm's frameworks and methodology. Maintenance mode. Hands free. They don't have to keep you on retainer, but they keep you in the building.
One firm described it to me as almost a leave-behind that keeps paying. The client gets a thing that remembers why every decision was made, and the firm gets a monthly cheque for work that's already done. When the next project comes up, guess who they call.
This is also the model that answers the "what happens after go-live" question that every implementation firm gets asked. Now you have a real answer.
A full-service demand gen agency we work with charges around $10,000 a month as their baseline. They get a lot of inbound from companies that want the help and can't pay that. Historically those were lost deals.
Now they say: think of this as the do-it-yourself version. For a few hundred dollars a month you get access to our brain, our playbooks, the way we think about your problem. You do the work, we're in your Slack when you get stuck. Last I checked they were charging somewhere in the $500 to $750 a month range depending on seats.
This turns "no" into a small "yes." Some of those clients grow into the full fee. The ones that don't are still paying you something for expertise you've already built, at close to zero marginal cost. If you have a steady stream of too-small-to-serve inbound, this is the model to look at first.
Some firms choose to give it away at first. They bundle it into the engagement as a value-add, don't charge for it, and treat the first few clients as a test. One founder's reasoning: "so that I look really smart in front of my customer." Fair enough.
To be clear, this is a choice, not a necessity. Firms can and do charge from day one. One firm we work with made $34,000 in its first 90 days selling access to its brain. They didn't run a free pilot, they put a price on it and sold it. If you have a clear use case and a client who already trusts you, there's no rule that says you have to earn the right to charge.
But if you'd rather learn first, the free route has real upside. The observability alone pays for it early on. You find out which questions clients actually ask, which use cases land, and what they'd pay for. You build the case study. Then you charge the next client.
The risk is that "for now" becomes forever. Clients who got it free will push back when you try to charge later. If you're giving it away, say up front that it's included for the first N weeks or the first phase, and that it's a paid thing after. Firms that set that expectation on day one have a much easier conversation on day 60.
It depends on how you already sell. If you sell projects with a clear end, start with model 3. If you sell retainers, start with model 1 and move toward a premium tier later. If you turn away small clients, model 4 is almost free money. And model 5 is optional. If you want proof before you put a number on it, run it with one or two clients. If you already have the proof, charge.
Now the part everyone actually wants: what do I charge?
I'm not going to give you a single number because there isn't one. But there is a way to think about it that gets you to a defensible number in about ten minutes.
Anchor to what they'd otherwise pay you. The brain replaces something. Usually it replaces the ad hoc questions clients send between meetings, the support hours after go-live, or the "can I just ask you something quick" calls that eat your afternoon. Price it against that. If a client would otherwise need five hours a month of your team's time to get those answers, and you bill $250 an hour, the brain is worth $1,250 a month to them before you've said a word about AI. Charge a fraction of that and it's an obvious yes.
Do not anchor to what the software costs you. Your client doesn't care what you pay for infrastructure any more than they care what you pay for your CRM. If you tell them it costs you $50 a month and you're charging $500, you've just started an argument you can't win. The value is in what's inside it, and what's inside it is your firm.
Flat and predictable beats usage-based, at least for now. You'll be tempted to price per query, per seat, or per something. Resist. The founder of a sales training firm, someone who knows what it costs to run these systems, put it well on a call: as long as he can predictably price it into the engagement, he's fine. What he worries about is usage costs blowing up mid-engagement and taking a low-margin project to a no-margin one. Your clients feel the same way about you. A flat monthly number they can budget for is easier to sell than a meter.
The big client problem (unsolved). Here's where it gets hard. One of our customers sold the same brain to a 100-person company and a 1,200-person company. The 1,200-person company has twelve times as many people who could use it. Should they pay twelve times as much?
I don't have a clean answer and I won't pretend to. What I've seen work is pricing by the size of the engagement rather than the size of the client. A bigger company usually means a bigger project, which already justifies a bigger number. If that doesn't hold, tier it by seats, roughly, in bands rather than per person. "Up to 10 users, up to 50, unlimited." Keep it to three bands. Nobody wants to count.
Numbers I've actually seen. For what it's worth, and these are early:
These will look low to some of you and high to others. That's fine. The right number is the one that's small relative to what the client already pays you and large enough that you'd notice if ten clients were on it.
Start higher than feels comfortable. It's easier to come down than to go up. If you give it away or price it at $99 and clients love it, you've anchored them at $99. If you price it at $750 and a client pushes back, you have room to move and you've learned something about the ceiling. Every firm I've seen underprice this regretted it within a quarter.
You've got the positioning and a number. Now the actual conversation. A few things that work.
Show it, briefly. Two minutes, not ten. Use a real example from another engagement with the details swapped out. An empty demo proves nothing. Say so: "This is from another project, details changed. Yours would be built on your work."
Then ask it something that matters in their world. Not a party trick. Something they'd actually send you an email about. When the answer comes back, point at where it came from: "That's pulled from the implementation plan and the call notes from March. Not the internet, not invented. Every answer links back to what it came from."
Let them watch it refuse. This is the move most people skip because showing your own product saying no feels backwards. Ask it what another phase would cost, and let it decline. Then say: "It won't quote you, won't commit us to dates, and won't say a word about our other clients. Anything commercial comes to a person here. That's deliberate."
This settles the two things every client is quietly worried about: that the AI will say something stupid on your behalf, and that their data is mixed in with everyone else's. Showing a refusal does more for trust than any security slide.
Frame the handoff by how you work. If you're project-based and leaving it behind:
"Normally when a project like this finishes, the deliverable stays and the reasoning goes. Eighteen months later someone asks why it's set up this way and the honest answer is the people who knew have moved on. That's the bit we've fixed. It stays live, it stays yours."
If you're retained, there's no natural end point, so the framing is compounding instead of handover:
"This gets better every month we work together. Everything we learn about your business goes into it as we go, so it's sharper a year from now than it is today."
Say the price when it comes up, not later. If you plan to charge for it after the project, say so the first time you mention it. Letting a client assume it's free and then invoicing them is how you lose the account. "It's included through go-live, and after that it's $X a month if you want to keep it" is a perfectly normal sentence. Say it early and it's a non-event.
Expect the security questions and welcome them. Where does our data live, who can see it, does it train on our stuff, can client A see client B's work. Have crisp answers. Ours: data is walled off per client, nothing trains shared models, access is controlled by the firm. If you can't answer these confidently, don't deploy yet. These questions are not objections. They're a sign the client is taking it seriously enough to care.
I'd be lying if I said every firm that tries this nails it. Here's where it breaks.
Clients don't change their habits. The founder of an agency operations consultancy named this one immediately: the challenge is on the end user shifting their habits to go to a new tool. He's right. If your brain lives in a separate portal with its own login, usage dies in week three. Nobody opens a new tab to ask a question they could email you.
The fix is to put it where they already are. Slack, Teams, whatever they live in. The best deployments I've seen are a bot in the client's existing channel. The client asks a question the same way they'd ask you, and it answers. No new habit required. This is the same gap McKinsey found across every industry: 88% of organizations say they use AI, but only 7% have it fully rolled out. The tool exists; the habit doesn't. Don't make your client build one.
Giving access to everything instead of one thing. When a firm first gets this working, the instinct is to open the whole brain to the client. Every framework, every doc, every call. Then the client doesn't know what to ask and asks nothing.
The firms that get adoption pick one use case first. "Ask it for a status update." "Ask it why we set up the pipeline this way." "Ask it what to do when X happens." One job, clearly named. Once the client is using it for that, they find the rest on their own.
Selling the tool instead of the firm. Covered above but worth repeating because it's the most common failure. The moment your pitch sounds like a software demo, you've lost. You're not competing with ChatGPT. You're offering more of the thing they already pay you for.
Underpricing and then being stuck. Also covered, also worth repeating. Free is fine as a deliberate first step with an end date. Free as a default because you couldn't decide is how this becomes a cost centre instead of a product line.
Deploying before the brain is good. A thin brain gives thin answers and the client's first impression is "this is useless." You don't get a second first impression. Before a client sees it, load it with real material: the project docs, the call transcripts, the decisions and the reasoning behind them. Test it yourself with the questions the client actually asks. If you're not a little impressed, they won't be either.
Nobody owns it internally. In a five-person firm everyone owns it, which means no one does. Someone needs to be responsible for keeping it fed, watching what clients ask, and turning that into the next thing to build. That's a few hours a week, not a full-time job, but it has to be someone's job.
If you've read this far you're probably thinking about which of your clients to try this with. Good. Here's the sequence that has worked.
That's the whole playbook. Pricing, packaging, positioning, and how to say it out loud. None of it is complicated once you stop thinking of it as software and start thinking of it as the first time you've been able to sell your expertise without selling your hours.
I'll be honest about one more thing. A year ago I would have told you the hard part was the technology. It isn't. We've built the thing, other people have built the thing, and it works. The hard part is the thirty seconds where a consultant who has only ever sold time has to say a monthly number for something that doesn't have any hours attached to it.
If you get that sentence out, the rest follows.
If you're a firm working through this right now, I'd like to hear what you've tried, especially what didn't work. That's the part of this playbook that's still being written.
