used AI to produce goods or deliver services in the previous 12 months
Canadian businesses, Q2 2026Live online · Toronto & the GTA
AI adoption is rising. Relevance is still the real barrier.
AI Complex is designing hands-on training for project managers, engineers and small teams that want to build AI-agent-assisted workflows around work they already do.
No broad AI lecture. No generic list of prompts. Start with one recurring task, build a working pilot, and keep people responsible for the final decision.
15 focused questions · about 4 minutes · contact details are optional
The relevance gap
Adoption is moving quickly. Many businesses still cannot see a useful first step.
These figures are Canada-wide benchmarks, not Toronto adoption rates. They help frame the question that our GTA research needs to answer.
reported the same two years earlier
Canadian businesses, Q2 2024said AI was not relevant to the goods or services they deliver
Most-cited barrier in Q2 2026Source scope: Statistics Canada, Canadian Survey on Business Conditions, second quarter of 2026. The survey asked about AI used to produce goods or deliver services during the preceding 12 months.
What we saw on the ground
A quiet inbox can still be a useful signal.
In our early outreach to GTA insurance brokers, accounting firms, property managers and law offices, response to a cold email was very low.
We do not treat that as a market statistic or a complaint. It is a working signal: small and mid-sized businesses are not waiting for another general AI pitch. The example has to be specific enough to their work to earn attention.
Our design rule: begin with one recurring task, show the human review point, and measure whether the result is actually useful.
Who the first programs are for
People who own a workflow—not people collecting AI tips.
The best starting point is a task that repeats, has a clear output, and still needs human judgment.
Project managers
Turn meetings, risks, decisions and status reporting into a controlled workflow with clear human approval.
Action logs · weekly status · risk summariesEngineers & technical leads
Structure requirements, change notes, technical research and test evidence without handing final judgment to a model.
Requirements · change review · technical briefsService professionals
Improve intake, document review and client follow-up in insurance, accounting, legal and property services.
Intake · document triage · follow-up draftsOperations & small teams
Create shared instructions, exception routes and quality checks so AI use is repeatable across a team.
SOPs · knowledge retrieval · handoffs
Hands-on curriculum
Build an agent-assisted workflow people can inspect, test and improve.
The course follows the same five-part build path whether the use case is a project update, technical brief, document intake or operating procedure.
Use approved, public, simulated or de-identified information. Employer policies and approved tools come first.
Choose the right task
Find one repeated, measurable and low-risk workflow.
Map the decisions
Define inputs, rules, exceptions, owners and the final output.
Build the agent-assisted flow
Create prompts, templates and tool steps around the real process.
Add evidence and review
Test quality, privacy, failure cases and human approval points.
Package the pilot
Leave with a reusable workflow pack and a 30-day rollout plan.
What paid participants leave with
Useful working assets, not a longer tool list.
Workflow map
Inputs, decisions, exceptions, owners and final output
Agent instructions
Reusable prompts, context rules and approved tool steps
Input and output templates
A consistent structure the team can review
Quality test set
Examples, expected results and failure cases
Human-review checklist
Fact, privacy, risk and approval checks
30-day pilot plan
Owner, baseline, measures and next decision
Free to understand · paid to build
Start with a useful example. Pay when you want guided implementation and finished assets.
Prices are first-round validation assumptions. They are not presented as GTA market averages and will be adjusted using survey and pilot conversion data.
AI Workflow Clinic
Free60 minutes · live online
See one complete workflow, learn how to choose a safe first use case, and decide whether a guided pilot is useful.
- One end-to-end workflow example
- Task-selection scorecard
- Privacy and human-review checklist
Workflow Agent Sprint
CAD 349Two weeks · four live sessions
Bring one approved task and build an implementation-ready, human-reviewed workflow with direct feedback.
- Four 2-hour working sessions
- Small group of 6–10 participants
- One finished workflow pack
Team Workflow Pilot
From CAD 4,500Up to six core participants
Design and test one shared workflow with ownership, access boundaries, quality checks and a team handoff plan.
- Workflow discovery and mapping
- Reusable templates and review rules
- Pilot evidence and adoption plan
A future four-week standard cohort at CAD 595 per participant will be considered only after the founding pilot is validated. Applicable HST is not included. Programs do not include production software integration, legal advice, job placement, income promises or guaranteed performance outcomes.
What this survey will decide
We are not asking whether AI sounds interesting.
We are testing which workflows matter, who owns the decision, what delivery format people will attend, and which price points are realistic.
Curriculum priorityWhich tasks and barriers come first
Audience fitIndividual, manager or team buyer
Delivery designFormat, schedule and start timing
Offer designFree clinic, paid sprint or team pilot
Before you join
Clear answers to common questions
Is everything remote?
Yes. Workshops, working sessions, feedback and office hours are delivered live online in Eastern Time. There is no physical Toronto classroom.
Do I need coding experience?
No for the first individual pilot. The focus is workflow design, agent instructions, testing and human review. Engineers may choose more technical use cases, but coding is not a general prerequisite.
Does “deploy an AI agent” mean a production system?
The individual sprint aims for an implementation-ready pilot around one approved workflow. Production integration, security review and organization-wide deployment are separate projects.
Can I use confidential company or client information?
Not by default. Use approved, public, simulated or de-identified information. Your organization’s policies, contracts and approved tools always take priority.
Who should not join?
This is not designed for someone seeking a certificate, a job guarantee, a fully autonomous decision-maker, or a general tour of every popular AI tool.
Shape the first GTA cohort
Start the survey →
Tell us which workflow is worth building first.
Your response will help set the curriculum, schedule, delivery format and pilot price. You can complete the survey without leaving any contact information.