The Amii AI adoption program has federal money to build AI projects with 25 more Prairie small businesses. The seats will pay off most for firms sitting on years of messy operating data.
$2.5 million across 25 firms is about $100,000 a seat on average.
Every success Amii lists started with a narrow, data-heavy problem.
A firm with an idea and no data history should wait.
The Amii AI adoption program now has $2.5 million more from Prairies Economic Development Canada to partner with 25 new small and mid-sized businesses in Alberta, Saskatchewan and Manitoba, according to Amii’s October 27 announcement. Our view: those seats are worth the most to firms with long, untidy operating records, and least to firms arriving with a product idea.
Key points
- Total federal support for the program is now $5.6 million.
- Firms keep the technology and IP, and Amii embeds ML residents to build in-house skills.
What is the Amii AI adoption program?
It is a hands-on program in which Amii scientists and embedded machine learning residents build custom AI systems with Prairie SMEs through what Amii calls Advanced Technology Projects. The business keeps the technology and the IP, and its own staff learn to run the system. The new $2.5 million from PrairiesCan funds 25 more of these partnerships.
In short: PrairiesCan’s new $2.5 million brings federal support to $5.6 million and adds 25 firms to the 30 Amii has worked with since 2024, which averages about $100,000 per seat. Source: Amii, October 27, 2026. The earlier-support bar is our subtraction ($5.6M minus $2.5M), and the per-firm averages are our division. Amii has not said the money is split evenly, and part of it funds running the program.
How much is each Amii seat worth?
On average, about $100,000 of new federal money per firm. That is $2.5 million divided by 25 businesses. Across the program, $5.6 million over 55 firms is roughly $101,800 each. However, these are our averages, not Amii’s budget lines, and part of the money will go to running the program rather than to any one project.
Next, the payoff. As an illustration only: staff spend 1,000 hours a year sorting field reports at a loaded $60 an hour, or $60,000. A model that removes 30% of that work saves $18,000 a year, so one seat’s public money equals about five and a half years of savings. Therefore, the system has to outlive the residency.
Which Prairie businesses should apply to the Amii AI adoption program?
Firms with years of operating data and one costly decision they make over and over should apply first. Every result Amii highlights, from forestry stream maps to gas well operations, started from a narrow problem with deep historical records. By contrast, firms whose AI plan is still a slide deck will get less from a seat than they expect.
Our ranking, best fit first, based only on the projects Amii describes:
- Field-data firms with a decade or more of records. Amii says 20 years of Carson Integrated’s field data became a stream classification tool with up to 90% predictive accuracy. This is the template.
- Operators of equipment that behaves unpredictably. The gas well project used reinforcement learning to cut costs and methane, by Amii’s account. Unread machine logs are raw material.
- Agriculture firms with imagery or yield maps. Similarly, the Croptimistic work added computer vision to crop maps that already existed.
- Software companies with a live product. Likewise, Visionstate and Riva added AI to platforms that already had customers.
- Firms starting from an idea. Last. Nothing in Amii’s examples suggests the program exists to decide what you should build.
What should a firm prepare before applying?
Before applying, a firm should be able to show where its data lives, how many years it covers, and the dollar cost of the decision it wants a model to improve. It should also name the employee who will work beside Amii’s resident and own the system afterward. Otherwise, the IP transfer has nowhere to land.
Even so, owners skip that last point. A model nobody inside the company understands becomes an expensive file once the residents leave, a problem ZAK lays out in his column on AI agent maintenance.
If the timing is wrong, compare AI funding in Canada that comes as a repayable loan and Alberta’s wider AI push.
Is 25 seats enough for the Prairies?
No, and it is not meant to be. Twenty-five partnerships cannot move adoption across three provinces on their own. Instead, their real value is as worked examples that other owners can copy. That is why the seats should go to firms most likely to produce a clear, measurable result, not spread thinly across firms still deciding what AI is for.
Admittedly, this favours established companies over younger ones. Still, we accept that. Twenty-five clear wins teach the next wave of owners more than 25 unfinished experiments. If idea-stage firms in the new cohort post results as strong as the field-data projects, we will change our ranking.
What to watch
- How fast the 25 seats fill. Amii says engagements are already underway in energy, agriculture and manufacturing.
- Whether results from the new cohort are reported in dollars. We would not expect them before mid-2027.
FAQ
Who can apply to the Amii AI adoption program?
Amii says the program is for small and medium-sized enterprises based in Alberta, Saskatchewan and Manitoba. The new PrairiesCan funding covers partnerships with 25 more of these businesses.
How much funding does the Amii AI adoption program have?
PrairiesCan added $2.5 million on October 27, 2026, bringing total federal support for the initiative to $5.6 million, according to Amii.
Does my business keep the IP from an Amii project?
Yes. Amii says its model includes complete technology and IP transfer, so the business owns and can maintain the system it co-builds with Amii’s team.
Our bottom line: bring Amii a decision you already make badly, at a cost you can name, with the records to prove it.
Fact-checked against Amii’s October 27, 2026 announcement on October 29, 2026. Dollar averages and the savings example are our own calculations.
