An AI workshop starts with one office task
A small pilot should leave the team with a repeatable method, a review routine and an example it can maintain.
An engaging workshop can still leave a team unsure how to use what it learned on Monday. We would begin with a task people recognize from their own projects, then carry it through preparation, review and handover.
Fraktal’s AI education moves through literacy, workflows, and agents and automation. A first pilot should connect the appropriate level to one concrete office need.
Turn a meeting note into a design brief
With permission to use the example, prepare a facade meeting note and its reference drawing. The note should distinguish agreed changes, open questions and suggestions. Participants read the source before reviewing an AI-assisted brief.
We would check whether fixed openings remain fixed, whether a suggestion became an instruction and whether an unsupported material choice appeared. The team needs a shared description of an acceptable result.
Try an unfamiliar example
OpenAI’s evaluation guidance emphasizes task-specific examples and human judgment. We would follow the prepared case with a second brief containing a missing dimension or conflicting instruction. That tests whether the method handles uncertainty beyond the rehearsal.
A quick first draft can still require extensive correction. Record preparation, review and correction effort together to understand where the method helps.
Leave something the office can reuse
- Source material cleared for use and an example of an accepted output.
- The instruction, review questions and a second practice case.
- A named maintainer, a version and a way to record failures.
Another team member should be able to repeat the task after the session. That is the outcome we would use to shape the first pilot.