Complete each sentence with ONE phrase from the box.
Ask the teacher: Write down anything you want to remember from this section. Your notes are saved automatically.
Rewrite each exaggerated sentence with an HONEST quantifier ('most', 'a few', 'hardly any', 'none') and — where relevant — name the exception.
Ask the teacher: Write down anything you want to remember from this section. Your notes are saved automatically.
Correct the ONE error in each sentence.
Ask the teacher: Write down anything you want to remember from this section. Your notes are saved automatically.
Decide whether each line is HONEST (uses quantifiers with a named exception / consequence) or ADOPTION-AS-SUCCESS (universal, unnumbered, unfalsifiable).
'Most client-facing decks — with a flagged carve-out for the pharma engagement — go through the copilot.'
'Every junior is using the tool — massive uplift across the board.'
'Hardly any drafts went out unreviewed — the ones that did are in the log.'
'All feedback has been positive.'
'A few Tier-1 clients have asked — I read it as the first three of many.'
'Everyone learned from the one client-facing error last month.'
Ask the teacher: Write down anything you want to remember from this section. Your notes are saved automatically.
Rewrite each hype line as a testable one-sentence policy that a manager would sign under.
Ask the teacher: Write down anything you want to remember from this section. Your notes are saved automatically.
You'll hear three partners brief the Managing Partner on their practice's copilot rollout. Decide which recommendation you'd sign under.
🎙️ Three partners brief the Managing Partner on the copilot rollout — multi-voice
Anya (Julian):Every junior on my team is using the copilot — literally every one. All our client-facing decks go through the tool first. Massive uplift across the board. Yes, there was that one memo last month, but it's a one-off. Everyone learned from it. Roll to 400.
Tiago (Anika):MOST client-facing decks — with a flagged carve-out for the pharma engagement — go through the copilot. A FEW Tier-1 clients have proactively asked; I read it as the leading edge. HARDLY ANY drafts went out unreviewed; the exceptions are in the log. My rule: NO copilot output leaves this firm to a client until a named manager-or-above has signed the log. Two client-facing corrections in a quarter and I pull the tool from that practice — that's the number.
Clara (Diego):In most cases we sort of review the output — most of the time. I think we should probably have some kind of sign-off. The risk is real, this can't just be a login. But I don't want to slow the team down. Maybe a light-touch review. We should probably discuss it more.
Ask the teacher: Write down anything you want to remember from this section. Your notes are saved automatically.
Write your OWN team's ONE-PAGE 'How we use AI' policy (max 130 words). Cover: (1) the human-in-the-loop rule, in one testable sentence; (2) the three monthly metrics (speed, quality, trust); (3) at least ONE quantifier ('most', 'a few', 'hardly any') used HONESTLY with a named exception. Sign it with a real role.
Ask the teacher: Write down anything you want to remember from this section. Your notes are saved automatically.