AI consulting pricing: what it costs, and how to know if it is worth it.

AI consulting pricing: what it costs, and how to know if it is worth it.

Most firms in this category will not put a number on it until the third meeting. We put one on it in the first thirty minutes, on the free call, once we know what the work is. This page is how the pricing works, the second bill that arrives after ours, and the arithmetic that says whether any of it is worth paying for. Sometimes it says no.

How each stage is priced

Four stages, each priced a different way. Select one to see what sets the number and what it buys. Yours is quoted on the call, against your operation, never read off a list.

Discovery: Free.

The one meeting where the number gets said out loud. Once we know what the work is, we tell you what it would cost, before any proposal and before you have committed to anything.

• Discovery call: Free, thirty minutes. If the honest answer at the end of it is that you should not buy anything from us yet, that is the answer you get.

Audit: A fixed fee, agreed before it starts.

Quoted before it starts, and it does not move. What sets it is the hours underneath: ten to thirty hours of a senior engineer sitting with the people who do the work, and it is the same engineer who has to build the thing afterward.

• AI readiness audit: Fixed fee, quoted before it starts. Two weeks mapping how the operation actually runs, five to ten places AI would take real work off your team, and a ranked plan with what each piece is worth. The plan is yours whoever builds it.

Build: A fixed fee, quoted against the audit.

The number is known once the work is. What moves it is how many systems the work has to cross and how much of your data has to be untangled first. Not hours: fixed scope, fixed fee, agreed at the end of the audit.

• One system: Fixed fee, one workflow. One workflow end to end, four to ten weeks against a fixed scope, with a working demo on your own data every week.
• Several systems: Fixed fee, priced by the systems it crosses. Work that has to cross between systems: the phone line into the platform, the platform into the CRM, the CRM into the reporting.
• Tool selection and implementation: Fixed fee, less than a build. When you do not need anything custom. Somebody picks the right off the shelf product, configures it properly, and connects it to what you already run.

Operate: A monthly fee, cancel any month.

Every custom build carries a setup fee and a monthly fee. A proposal with a setup fee and no monthly fee has not answered what happens when the software on the other end changes.

• Voice agents: A setup fee, then a flat monthly fee. Our most requested build and the one with the clearest before and after. The phone answered at any hour, triaged by your rules, written into your systems.
• Search and content: A monthly fee, scoped to the work. A wide band because the work is genuinely different at each end of it, from keeping structured data honest to writing the pages that earn the search.
• Ongoing operation: A monthly fee, cancel any month. Not waiting for something to break. Somebody continuing to build, watching what is running, and shipping something new each month.

There is a bill after ours

Our fee is for the work. The system we build for you runs on software that bills you separately, every month, for as long as it runs.

The number most proposals leave out, including, historically, some of ours.

Depending on what gets built, that can include an automation platform, a model API, a telephony provider, a database, and hosting. None of it is enormous. All of it is real, and it does not stop when the build ends.

So the monthly fee on a proposal is never the monthly cost. It is that fee plus the running costs underneath it, and whoever is selling it to you should be able to tell you the second number before you sign. We quote it separately, in writing. Ask any other firm for the same thing.

The monthly cost, in full. Who bills you:

• Our monthly fee: Us. On the proposal, in writing, before you sign.
• Automation platform: The vendor.
• Model API: The vendor.
• Telephony: The vendor.
• Database and hosting: The vendor.

Which lines apply depends on what gets built. We put the second number in writing before you sign. Ask any other firm for the same thing.

Three tests, one of which you can run right now

Three tests, in the order they are worth running. The first one is two sliders and a minute of your time.

• The volume test. Take the monthly fee, divide it by the profit on one closed job, and you have the number of jobs the system has to save you a month. Then ask whether you lost that many last month to a call nobody answered or a quote nobody chased. If the answer is comfortably yes, the spend is arithmetic rather than faith. If it is close, wait. Most operations cannot run this test at all, which is itself the finding: nobody is measuring the thing the system would fix.
• The leak test. The biggest return in a small business deployment is usually money that was already leaking and nobody was counting: the after hours calls going to voicemail, the leads that arrived on a Saturday, the invoice that got paid twice. You cannot price that before the build, because the whole reason it is leaking is that it is unmeasured. What you can ask for is a system whose first job is to make the leak visible, and a date by which you will see it.
• The hours test, which is the weakest one. Hours given back are real, and they are the return people quote most often. Ten hours a week off a coordinator, multiplied by a rate, divided by the fee, and out comes a percentage. That is arithmetic, not return. Recovered hours turn into money only if the person spends them on something that earns. If the honest answer is that the ten hours become a calmer week rather than more revenue, that is a fine reason to buy and a bad reason to expect a payback number.

Did you lose that many last month to a call nobody answered or a quote nobody chased?

• Comfortably, yes. Then the spend is arithmetic rather than faith. Bring the number to the call and we will quote against it.
• It is close. Then wait. Count for another month and run this again. A system that has to work perfectly to pay for itself is a bet, and we would rather not sell you one.
• I cannot tell. That is itself the finding: nobody is measuring the thing the system would fix. A month of counting missed calls and unchased quotes costs nothing, and it is the first thing we would ask for anyway.

Four times the answer is no.

Four questions we ask on the free call. Answer them here first, and if one comes back no, you have saved us both a meeting.

We turn down work for each of these, and it is cheaper for both of us to find out here than on the call.

• You cannot name the work. Can someone on your team describe the task in one sentence? If nobody can describe the task in a sentence, there is nothing to automate yet. What you need is a month of writing down what actually happens, which costs nothing.
• The volume is not there. Does it come up more than a couple of times a week? A system that saves ninety seconds on something you do twice a week is a hobby. Below the line the right spend is an off the shelf tool and an afternoon, and we will tell you so on the call.
• Nobody will own it. Is there one person inside the business who will care whether it works? Every system that survived contact with a real operation had one person inside the business who cared whether it worked. Without that person, the build lands and quietly stops being used.
• The goal is removing a person. Will the people doing this work today still be there once the system runs? We do not take those engagements. The routine work goes to the system and the judgment stays with your team. If what you want is a smaller payroll, we are the wrong firm.

What a cheap quote is buying you.

We lose deals on price, and some of those are the right outcome. These are the ones that go wrong.

The cheap quotes that go wrong go wrong in a predictable way. Someone builds the happy path, it works in the demo, and then it meets your actual inbox.

A build is typically ten to thirty hours of senior engineering time. Divide any quote you get, ours included, by a plausible number of hours and see whether the result is believable. Then ask these five, of every firm on your list:

• How many hours of engineering time is this, and who is doing them?
• What happens to my system when the software on the other end changes?
• What is my monthly running cost, separate from your fee?
• Can my team turn it off without calling you?
• What is the plan for the cases it gets wrong?

A firm that cannot answer the fifth one has thought about the model rather than about your operation.

Asked once the numbers are on the table

The questions that come up once the numbers are on the table.

Is hiring an AI consultant worth it? It is worth it when you can name the work, the volume is there, and somebody inside the business will own the result. Test it yourself first: take the monthly fee, divide by the profit on one closed job, and ask whether you lost that many jobs last month to a call nobody answered. If the answer is comfortably yes, the spend is arithmetic. If you cannot run that test at all, the honest first step is measuring for a month, which costs nothing.

How much does an AI consultant cost? It depends on which of four things you are buying, and we tell you which on the free discovery call. The audit is a fixed fee agreed before it starts. A build is a fixed fee quoted against the audit, so the number is known once the work is. Tool selection is cheaper than a build because nothing gets written from scratch. Operating what we built is a monthly fee you can cancel any month. What drives all four is scope, how many systems the work has to cross, and how much of your data has to be untangled first. What the wider Toronto market charges, and how to read a quote against it, is in the market survey.

What does it cost to implement AI in a small business? For most small operations the first real system is a single build plus a monthly fee, and the first thing built is usually the phone line. If you do not need anything custom, tool selection and implementation costs less than a build and gets you a properly configured off the shelf product instead. Either way the fee sits on top of the software the system runs on, which is billed to you separately and quoted before you sign. The arithmetic for whether it is worth it at your size is in can a small business afford an AI consultant.

How much do AI voice agents cost? A setup fee, then a flat monthly fee, plus the telephony and model costs the agent runs on. Flat matters: the human answering services meter by the minute or by the call, so their bill rises exactly when your business gets busy. It is the most requested build we do and the easiest one to test before you buy. Take the monthly fee we quote you, divide it by the profit on one job, and you have the number of missed calls a month the agent has to convert to pay for itself.

How much does an AI readiness audit cost? A fixed fee, agreed before it starts, and you hear it on the free call. What sets it is the work under it: ten to thirty hours of a senior engineer's time, and it is the same engineer who has to build the thing afterward. Most of an audit is not technical. It is sitting with the person who does the work and watching what they actually do, because what people describe and what people do are different processes. Take whatever any firm quotes you, divide it by a plausible number of hours, and see what you are actually buying. A quote that implies two or three hours is selling you a document.

Why is there a monthly fee at all? Because something on the other end always changes: a platform ships an update, a phone menu moves, a form field gets renamed. The monthly fee covers watching what is running, fixing what moved before you notice it, and shipping something new each month. A build sold as set and forget is the one that fails quietly six weeks later.

Do you charge hourly? No. Fixed scope, fixed fee, agreed at the end of the audit when the work is actually known. Hourly billing puts our interest and yours on opposite sides of every decision, and a build priced on top of a real map of your operation does not need it.

What if the audit says we should not build anything? Then that is the deliverable, and you still keep the plan. It has happened, and it is a better outcome than a build nobody uses. If you take the audit to another firm, you will get a better quote from them too, because they will be quoting against a map instead of a conversation.

Can you give me an ROI percentage? No, and be careful with anyone who does before the build. Every percentage in this category is back-calculated: hours saved, multiplied by a rate somebody chose, divided by a fee, rounded to something persuasive. What we will commit to is an observable outcome and a date. The longer argument is in why we will not give you a percentage.

The four longer reads behind this page: what AI consulting actually costs, what the Toronto market charges, can a small business afford this, and why we refuse to quote an ROI percentage. Run your own missed call math in the ROI calculator, see what the money buys on the services page, or start with the questions owners ask first. If the question is how to pay for it, AI funding in Canada covers LIFT, the Regional AI Initiative, and what happened to CDAP.