5 signs your team is spending too much time on repetitive questions

Nobody decides to spend their week answering “what time do you close”. It accrues. Volume grows, the same forty questions grow with it, and at some point the team’s day is mostly typing things the business already knows.

The difficulty is that it rarely announces itself. Here are five things that show up first — each with a way to check it this week, without buying anything.

1. Your saved replies have saved replies

Every team builds canned responses. The signal is not that they exist; it is when someone starts organising them — folders, naming conventions, a doc explaining which to use.

At that point you have built a small, manual, human-powered answering machine. It works, and it costs a salary.

Check it: open your saved replies and count them. Then count how many were used this week. If the top five cover most of the volume, those five are the automation case, and you already know exactly what they should say.

2. New starters are productive in a day

Onboarding speed is usually a good thing. For an inbox it is diagnostic: if someone with no product knowledge can handle the queue by their second day, the queue is not asking for product knowledge.

The work has become lookup, and lookup is the thing software is unambiguously better at.

Check it: ask whoever onboarded last what they actually needed to learn. If the honest answer is “where to find the price list”, that is your answer.

3. Your best person is your fastest typer

Watch who gets praised. If it is throughput rather than outcomes — who cleared the most conversations, not who saved the most deals — the job has quietly been redefined as clerical.

This one has a cost beyond time. People hired to solve problems and given a typing job leave, and they leave with the knowledge that made them good at the hard conversations.

Check it: look at your last two departures from the team. What did they say in the exit conversation?

4. Evenings and weekends look nothing like weekdays

Pull response times by hour. Most consumer businesses find a cliff at 6pm: messages keep arriving, answers stop.

Those out-of-hours messages are overwhelmingly the repetitive kind — hours, availability, price, do you do X — because that is what people ask when they are deciding, and deciding happens in the evening.

This is the clearest case of all, because it is not fixable by working harder. A rota that covers 8pm on a Sunday costs more than the conversations are worth, for questions a machine can answer correctly.

Check it: count the messages that arrive after your last person logs off, and how many are one of your top five saved replies.

5. Nobody knows what people actually ask

Ask three people on the team for the top ten questions. If you get three different lists, the repetitive work is invisible as well as expensive — and invisible work cannot be prioritised, staffed, or automated.

Check it: take the last 100 conversations and tag each with one word for the intent. It takes about an hour. The distribution is almost always more concentrated than anyone expects, and it is the single most useful hour you can spend before buying any tool.

What to do with the answer

If three or more of these are true, the question is no longer whether to automate the repetitive half — it is which part, and with what guardrails.

The order that works:

1. Automate the top five intents only. Not forty. Five, from the tagging exercise above, where the answer is a fact you already maintain.
2. Make the source the real source. The agent should answer from your price list and your calendar, not a copy that drifts.
3. Escalate everything else, with context. The measure of a good deployment is not how much it answers, it is how cleanly it gives up.
4. Watch the failures, not the successes. The questions it could not answer are your roadmap for the next month.

The goal is not an inbox without people. It is an inbox where the people are doing the part that needed a person.

---

Fellix answers the repetitive half from your own information, and escalates the rest with the whole conversation attached. See how the Front Desk handles it.