PracticeWorking with AI: Delegation & Oversight

~2–3 minChoose, then see feedback · no writing

Choose the check, then the rollout

The 96% green light

The situation

Tidewater sells helpdesk software to business customers. An AI model drafts Tidewater's own support replies, and today a person checks every draft before it is sent. The model was tested offline on past emails. Maya, the product lead, wants to know which replies, if any, can go out without that check.

Maya · Product lead: “We have one analysis slot. What do you need before making the call?”

You can request one analysis before tomorrow’s planning meeting. Choose what to inspect, then recommend what the team does tomorrow.

What you can see (3) · tap a card to read it

›E01 · Headline

An offline test used 200 historical English email requests from existing customers. Reviewers judged 192 proposed replies acceptable. No test reply was sent to a customer. “Acceptable” means correct issue identification and a supported response with no prohibited account change or unsupported promise.

›E03 · Boundaries

The test excluded live chat, multilingual requests and a forthcoming new product channel.

›E06 · Capacity

A reviewer can run a 20-reply trial tomorrow without delaying the current queue, timing each review and logging the request type and every edit, which today's workflow doesn't record. All 20 will be checked before sending. The team could instead wait two days for a broader offline analysis, which delays those measurements.

Step 1 of 2

Which one analysis do you request?

Your first choice is kept. Changing your mind later counts as a retry.