Which Accounts Haven't Heard Back in Seven Days? Ask for the List, Not the Mood

"How are we doing?" is a question that feels productive and rarely is. The answer comes back as a paragraph of reassurance: volume is steady, a few things are slipping, the team is on top of most of it. Nobody can act on that paragraph. You finish reading it and still don't know which customer to call before lunch.

The fix is not a better tone from your AI assistant. It is a sharper question. Compare the vague version with this one: "Which accounts had no reply to a ticket in the last 7 days?" The second question names a population (accounts), a condition (no reply), and a window (seven days). Any system that can read your records can answer it, and the answer is a list you can work through.

Why specificity changes the output

A language model answering an open question has to decide what "doing well" means. It picks a reasonable interpretation, usually an average or a trend, and summarizes that. The result sounds confident and hides every decision it made along the way.

A question with a clear definition removes most of those decisions. "No reply" has a fairly firm meaning in a ticketing system: there is an inbound message, and there is no outbound message from your team after it, within the window. The assistant is not guessing at sentiment. It is filtering records and returning the ones that match.

This is also why the output is checkable. If the list includes an account you know was answered, you can open that ticket and see why the filter missed it. A vague summary offers no such handle. You cannot audit "things look mostly fine."

Writing a question that returns a list

A few habits make the difference:

Name the thing you want to act on. "Accounts," "tickets," "invoices," and "deals" give the assistant something to return. "Performance" and "health" do not.

Put the condition in measurable terms. "Had no reply" beats "is neglected." "Renewal date within 30 days" beats "coming up soon."

Set the time window explicitly. "In the last 7 days" removes the ambiguity about whether you mean this week, the last sprint, or since the start of the quarter.

Ask for the fields you need to act. "Account name, ticket title, days since last inbound message, and owner" turns a list of matches into a call sheet.

Ask for a count only when you need one. A count is a summary, and summaries are where vagueness creeps back in. If you want the number, ask for the list and count it yourself, or ask for both and check that they agree.

What to do when the list is long

Sometimes the answer is forty accounts, and that is useful information too. Do not reply by asking the assistant to narrow it into a feeling. Narrow the question instead: add a plan tier, exclude accounts with an open escalation, or limit it to a single owner. Each refinement is still a filter, so each result is still a list.

The goal is a conversation where every answer ends with something you can do before the next meeting. Vague questions end with a shrug. Specific ones end with a name, a date, and a person responsible for the next step.