AI with Michal

Team workshops

Workshop on Personal Productivity with AI

Private cohort workshop for managers, directors, and high-performing ICs. These pages spell out format, sample exercises, tooling, logistics, optional post-workshop engineering, and how to book.

All team workshopsAbout me

Agenda & samples

Example agenda and sample exercises

Timings shift with cohort size and depth. Most of the time is spent rebuilding real artifacts (anonymized) and writing guardrails your team can reuse next week.

Modules (typical half-day / full-day)

  1. 1

    Operating model: draft versus decide

    ~20 to 30 min

    Outcome: Shared vocabulary for assisted steps versus steps that need a named reviewer before anything external.

  2. 2

    Inbox and calendar triage without theater

    ~30 to 45 min

    Outcome: Labels, defer templates, and a three-line rule for when a meeting actually needs a live brief.

  3. 3

    Meeting briefs people read

    ~25 to 40 min

    Outcome: One-page briefs with decisions, owners, unknowns, and live questions that respect redaction rules.

  4. 4

    Doc and memo summaries that keep unknowns

    ~25 to 40 min

    Outcome: Structured summaries that never invent numbers and call out gaps managers can act on.

  5. 5

    Rewrite loops for mail and chat

    ~30 to 45 min

    Outcome: Two variants plus a policy pass list before send on sensitive topics.

  6. 6

    Manager modeling and rollout checklist

    ~20 to 30 min

    Outcome: Exemplars directors can reuse and a dated checklist for the first two weeks after training.

  7. 7

    Team playbook handoff

    ~15 to 25 min

    Outcome: What to standardize Monday, what to measure lightly, what to revisit in 30 days.

Sample moments (marketing-safe)

Prompt frames are teaching scaffolding. Your team adapts them to your policies and templates. Do not paste regulated or customer-identifiable data into a live classroom unless your security team signs off.

Meeting brief from a noisy thread

Context

A director lands in ten minutes and the calendar invite has six attachments and forty-three messages. You need a one-page brief: decisions needed, risks, and three questions to ask live.

Prompt frame (template)

Given [paste redacted thread summary or bullets], produce: (1) five-bullet exec brief, (2) decisions needed with suggested owner, (3) three live questions, (4) explicit unknowns. Do not invent facts not in the input.

Quality checks

  • Unknowns are explicit when the thread is thin.
  • Owners are roles, not blame.
  • Questions are answerable in the room.

Before

Forwarding the whole chain and saying read below.

After

A short brief the director can scan in two minutes with clear asks and visible gaps so the meeting stays on decisions.

Inbox triage plus reply variants

Context

High performers drown in mail that needs fast, accurate tone. Draft two reply variants plus a defer template, then flag anything that needs Legal or HR eyes.

Prompt frame (template)

Given [redacted email thread], produce: (A) triage label (FYI / decide / schedule), (B) reply variant direct, (C) reply variant softer, (D) defer template under 40 words, (E) policy flags if any claim is unverifiable.

Quality checks

  • No invented commitments or dates.
  • Flags call out sensitive topics even if tone looks fine.
  • Defer template names the next concrete ask.

Before

A long draft that buries the ask and copies legal risk from the thread.

After

Two tight variants, a defer path, and a visible flag list before anything sends.

Doc summary with unknowns surfaced

Context

A twenty-page internal memo must become a manager read for Monday standups. Accuracy matters more than speed.

Prompt frame (template)

Summarize [redacted doc text or outline] into: (1) purpose in one sentence, (2) five bullets of substance, (3) decisions implied, (4) unknowns or missing data called out, (5) recommended next step. Do not fabricate numbers.

Quality checks

  • Every bullet ties to a section or says gap.
  • Decisions are labeled as implied versus stated.
  • No invented metrics.

Before

A confident wall of text that hides what is not in the doc.

After

A short read with explicit gaps so managers do not act on confident fiction.

Weekly personal AI retro

Context

Directors want habits, not heroics. Individuals log what they tried with assistants for a week, then compress into a five-line retro with one change for next week.

Prompt frame (template)

Given [bullet log of attempts], produce: wins, misses, one habit to keep, one habit to stop, one experiment for next week with a five-minute time box.

Quality checks

  • Experiments fit real calendars.
  • Stops name a behavior, not a person.
  • No vanity metrics.

Before

Vague I used ChatGPT more this week with no behavior change.

After

A dated retro line managers can scan in a one-on-one without shaming individuals.

Manager rollout checklist

Context

After the session, managers need a one-page checklist to roll norms out: approved surfaces, example prompts, and what to measure lightly for two weeks.

Prompt frame (template)

Produce a manager rollout checklist for [team type] with sections: approved tools, banned pastes, three exemplar prompts, weekly office-hours plan, and two lightweight metrics that are not vanity.

Quality checks

  • Metrics are behaviors or minutes, not mystery scores.
  • Banned pastes reference HR or Security guidance placeholders.
  • Exemplar prompts are short enough to reuse Monday.

Before

A slide titled AI policy with no actions.

After

A checklist with owners, dates, and exemplars people can run without another deck.