Do you cut people with AI?

That’s not the pitch. The pitch is the team you already have getting dramatically more done. As roles reshape we handle it the humane way: retrain, reallocate, don’t replace leavers where the work has moved on. Deflate the balloon, don’t pop it. Fewer chairs over time, maybe. An ambush, never.

What is the shadow AI audit?

We find where AI is already being used inside your company without official approval: the tools, the data going into them, and the duplicated spend. You get a map of the exposure and a plan. It is the fastest way to see the real picture, and a good first engagement.

How does monthly maintenance work?

After the build (fixed fee, quoted up front), care is a flat monthly fee with a minimum term: monitoring, fixes before you notice them, model and API updates handled. The minimum term exists because a system nobody maintains rots. It protects your investment, not ours.

Which AI models do you use?

Whichever fits the task. We are an independent third party, not tied to any vendor, so we match the model to the job instead of routing everything through the most expensive one. We are planning a “right model, right task” series to explain these choices in the open.

Do you work in person?

Happily. Remote by default, and remote works, but in person is more effective, we find. You can’t see how work really happens over a video call: the interruptions, the workarounds, the things nobody thinks to mention. It costs more (a day rate plus travel and hotel at cost and a flat daily per diem, quoted before you commit), and for the Map especially, it’s worth it. We’re based in Toronto and happily work on-site anywhere in the world.

What happens to our data?

It stays yours. We work under NDA whenever you want one, keep access to the minimum needed, and never reuse your data or show your business to anyone else. Confidential by habit.

We're not technical. Is this for us?

Yes, most people we talk to aren’t. We speak plain English, and the Map is drawn with you, not at you. If we can’t explain it in plain words, we don’t understand it well enough to build it.

Fixed quotes, never hourly. The Map is free. The diagnostic is $250–$500, credited in full if you build. Builds are quoted up front ($2k–$12k for a typical workflow) and the price on the quote is the price you pay.

Right model, right task: a planned comparison series

This is a planned series, not a published scorecard. This planned series will compare the practical AI tasks businesses use (drafting emails, summarising documents, generating images, transcribing calls, coding help, data extraction, and more), and we match each one to the model that does it well without overpaying.

The principle is simple: as an independent third party, we have no vendor to favour, so we judge models the way a good product manager judges features: fitness for the task, not the size of the price tag.

Microsoft’s small models, frontier models, open-weight models: each has a job it wins at. Our scorecard will say which, and why. If a task needs the premium model, we’ll say that too. Honesty is the whole point.

Smooth AI Consultancy homepage

Book a free working session