Meeting notes to file, follow-ups to write and information to move between tools: I build custom AI agents and agentic workflows for teams in Montréal and Toronto, then keep them running after launch.
A useful AI system needs to know how work gets done in your company: where the information lives, who needs it next, and which exceptions matter. I start with one recurring workflow, connect the tools involved, and test the system against your own work.
The system can read the relevant records, prepare a draft, update a tool or route a task to the right person. You keep the decisions that need human judgment, with approval steps agreed before launch.
A system connected to 27 tools drafts meeting records, weekly digests and follow-ups in each person's voice. Separate review steps check the writing before it goes out. I built it, host it and maintain it.
See the chief of staff systemA recruiting firm's Teams transcripts become structured notes in French or English, matched to the right candidate or client. The system updates empty profile fields and follows up with the consultant for their rating.
See the recruiting workflowFor a biotech company, an intake form starts the offer draft, approval route, DocuSign preparation, account setup and first-day calendar. The workflow connects HR, IT and the hiring manager's work.
See the onboarding workflowI keep the system running, check its outputs and update it when your tools or process change. Your team gets a record of what it handled and where a person stepped in. Scope and price are agreed before the work starts.
Read about security and dataI work with founders and operators at startups, small professional firms, and venture funds supporting their portfolio companies. Projects can run remotely, in English or French, around the tools your team already uses.
If your team wants to build its own systems, hands-on workshops start from the work in front of you. Advisory is another way to decide which workflow to tackle first and what it would take to run it reliably.
An agentic workflow uses AI to read context, choose the next step within a defined scope, and work across your tools. For example, an interview transcript can become a structured note on the right CRM record, followed by a task for the consultant. A person keeps the decisions that require their judgment.
Yes, Unsupervised is based in Montréal and works with teams in Montréal, Toronto and across Canada, in English and French. Work can be delivered remotely, with the format agreed before the engagement starts.
Yes, the systems I build connect to the tools your team already uses. Existing examples include Microsoft Teams, Google Workspace, Slack, Notion, recruiting CRMs, Rippling, Lattice and DocuSign. The connections and permissions are scoped to your workflow.
I host and maintain the systems I run for clients, test their outputs, and update them when the process or tools change. Your team can see what the system handled and where a person needs to step in. The build and ongoing service are scoped before work starts.
Bring one example of the recurring work and the tools involved, and we can start there.