AI your people can actually use

A practical keynote on tools, judgment, and work that actually ships. Not a tool tour. Not “10 prompts for Excel.”

Teaching spine:Feed itAsk itStill check

Companies new to AI — or stuck after buying seats.

Working session on their jobs: pick a decision → build the use-case → run it → break it → write failure modes.

Outcomes we hold to: (1) Someone can run this next week without us. (2) Someone can attempt a second use-case with the same pattern.

We teach strengths and weaknesses. Leadership-only vs company-wide is a strategy choice — we recommend one; we will not sell the larger option just because it is larger.

Five modules

Teaching spine:Feed itAsk itStill check

Why seats ≠ use — and why a launch email is not training.

Edges, not monopolies. Grok Bot as a practical home base; when to open other tools.

Monthly-pass feel vs. per-doc myth. Ballpark only — reconfirm prices at checkout.

Concrete department jobs and a use-case library you can reuse.

Red zones, guardrails, PC/mobile setup, two-week experiment.

Prompt pattern: Role · Goal · Input · Rules · Output · Check — plus “If evidence not present, say so.”

Harborline scenarios in the live keynote are fictional teaching composites — methods only, never presented as client work.

How the room runs

Leadership room

Decision-makers map one workflow and choose scope — leadership-only vs broader.

Two-week experiment

One job, baseline, review, refine, decide keep / change / stop. Measure usefulness, not enthusiasm. Start with one job. Build trust by checking the work.

Same method, your industry's paperwork.

Each edition is the same half-day working session: map one real job, run it on the documents your team already works from, try to break it, and write the check step. People send, people sign.

Request an edition →

Bring Practical AI to your team

Book a Practical AI workshop — or a 90-minute working session on one real workflow.