What "AI consulting" should mean if you're searching for it in Kelowna

The word consulting hides two different jobs

Somebody typing "AI consulting Kelowna" into Google at 9pm is usually not looking for a workshop. They have a specific process, a quote sheet that gets re-typed into a scheduling system, a compliance binder assembled from four inboxes every month, and they want to know if a person can go build against it. Most of what gets returned for that search is a strategy session. That is not the same job.

A strategy session ends in a document. Somebody who has spent three years watching a controller re-key numbers between a quoting tool and a job-costing spreadsheet does not need a document. They need the two systems to talk, or the quote to generate the job automatically, or the compliance pack to build itself from the four places it currently lives. That is an integration job, built against one business's own data and its own software, not a general recommendation that would apply to any business in the trade.

This search is rare, and it produced MannVenture's first customer. Keyword tools report almost no volume for it. The person searching it already knew what they wanted; they just didn't know the plain-English name for it yet. Custom AI is our name for that job: a system built against a business's own systems and data, published as a monthly engagement at $1,395 CAD a month rather than priced by the hour or the day.

Why a monthly price and not a project fee

One-time builds exist in this market and they are a real option: published prices for a one-time agency build in this market start around $7,000 (checked 22 September 2026), and you own the code and the accounts outright when it's done. The reason we don't sell Custom AI that way is what happens after the invoice clears.

A system built against a company's live systems has an ongoing relationship with software that keeps changing under it. Quoting tools update their API. Scheduling platforms change how they authenticate. A model provider changes pricing or deprecates a version. None of that is dramatic on its own, but each one breaks something quietly if nobody is watching for it, and a $7,000 build with nobody on retainer to notice the break sits broken until someone happens to check.

That's the argument for a monthly number instead of a lump sum: it buys somebody whose job is to notice. It's a smaller number every month rather than one larger number up front, and it keeps running because someone is paid to keep it running. A one-time build with no one watching means the business owner becomes the one who has to notice, and most controllers running a 15-person operation don't have that on their calendar.

What the model costs, and why that's not the number to worry about

Model usage on a running system typically runs $40 to $180 a month in provider fees, billed separately by the provider, according to the same published agency prices. That is small next to the $1,395 monthly engagement, because most of what a Custom AI build costs is the work around the model.

The model call itself is close to free at the volume a 10-to-50-person operation generates. A quote request, a scheduling lookup, a compliance summary; these are cheap operations for a model to perform. What costs money is everything around that call: understanding the business's actual quoting logic well enough to automate it correctly, building the connection into whatever software the business already runs (which is rarely documented anywhere), testing it against real edge cases like a job with three change orders, and then being present when it breaks.

Where the packaged products already cover it, and where they don't

Before anyone books a Custom AI conversation, check whether the problem is one of the three products already built. A slow reply to web inquiries is the AI Website. A phone that goes unanswered while hands are full is the AI Receptionist. Marketing that stops the moment the owner gets busy is the AI Growth Engine. None of those need custom work.

Custom AI starts where those three stop: a business's own data, its own systems, its own specific process that no packaged product was built to fit. A physio clinic with two locations pulling patient scheduling from one system and billing from another is a Custom AI conversation. A physio clinic whose website doesn't answer questions after hours is not; that's the AI Website at $97 CAD a month, already built, already live in days rather than weeks.

The search "AI consulting Kelowna" often turns out to be a request for one of the packaged products. Somebody asking for consulting because their phone goes unanswered on job sites doesn't need a monthly engagement built around their internal data; they need the receptionist, which already exists at $97 CAD a month against $1,395 for Custom AI. Getting that sorted on the first call saves everyone the wrong invoice.

Where Custom AI is the wrong answer

A business with under ten staff, one location, and no software of its own to integrate against almost never needs Custom AI, because there is nothing custom to build and nothing for the monthly fee to run.

The clearest sign a business isn't ready is that its "process" lives entirely in one person's head or one spreadsheet, not across two or more pieces of software that need to be connected. If a Chilliwack flooring company's quoting process is one estimator writing numbers into a template and emailing it, there's no integration to build; the fix might be the AI Growth Engine or a better estimating tool, not a custom system. Custom AI is worth $1,395 a month when there are two systems that don't talk to each other and a person paid to be the translator between them.

The other wrong-fit case is volume. A paid consultant's audit (around $585 CAD in this market) buys a document and an analysis. Our free AI audit is a faster first step for someone who doesn't know yet whether the manual process is worth automating. If the answer is "maybe two hours a month," a monthly engagement at $1,395 doesn't make arithmetic sense against the labor it replaces. What AI costs a small business walks through that arithmetic in more detail.

What to expect if you call

A conversation about Custom AI should end with a concrete answer about the specific system, not a proposal to "explore possibilities." The published price is $1,395 CAD a month. Anyone quoting a day rate or a block of hours for this kind of work isn't pricing it the way we do.

The test for whether a consulting conversation is worth having is whether the person on the other end can name the specific system within the first fifteen minutes: which two pieces of software need to connect, what data moves between them, and roughly what breaks first when one of them updates. If the conversation stays at the level of "AI strategy" or "digital transformation," that's a sign the work being sold is a strategy document rather than a working system.

For a 10-to-50-person operation in Kelowna, West Kelowna, Vernon or Penticton with one real bottleneck, that bottleneck is usually specific enough to name on the first call: a quote that gets re-keyed, a scheduler doing by hand what a rule could do, a compliance pack pulled from four places every month. Naming it is most of the qualification work. The rest is whether $1,395 a month against that labor cost makes sense, which is the arithmetic in AI vs. hiring.