Reviews and phone calls hold the most honest feedback you get, and almost nobody has time to read all of it. Your AI assistant can. Ask it to read both, group what it finds into themes, count each one, and quote a real example, so you know what to fix first.

See it in action

The AI Assistants page in TripWorks shows this request as a sample. The themes, counts, and quote below are sample data, not a real account.

Try the prompt with your own time frame:

Pull customer complaints from reviews and phone calls over the last 90 days. Group them into themes, count each, and quote the worst example of the top one.

What your assistant can read

• Reviews. The star rating, the written comments, and what the guest said could be improved. Reviews come from your guest review form and from imported sources such as Google, TripAdvisor, and Yelp, and each is tied to the activity it's about. TripWorks also scores each review's sentiment (positive, negative, neutral, or mixed), so your assistant can start with the harshest ones.
• Phone calls. The summary and transcript of each call TripWorks recorded, including AI callbot calls, plus the questions and tips found in the calls. Your assistant can also count calls by day, by how they ended, and by average length.

Your assistant does the grouping itself, so the themes reflect what it reads, not a fixed list.

Example requests

• Summarize this month's reviews. Gives the average rating, what guests love, and what keeps coming up.
• List every review under four stars from the last 60 days, newest first, with a suggested reply. You get a to-do list with a draft response for each.
• Rank my activities by average review rating over the last 90 days. Shows what to promote and what to fix.
• Find reviews that mention parking. Searches the comment text and quotes each match.
• What are callers asking about most this month? Reads the questions found in recent calls and summarizes the themes.
• Summarize yesterday's call from the number ending in 4471. Tells you what the caller wanted and what happened, without listening to the recording.
• How many voicemails did we get last week, and on which days? Counts voicemails by day so you can see when callers can't reach anyone.
• Which complaints show up in both reviews and calls? Finds themes that repeat across both, which are the ones guests feel most strongly about.

Turn what you learn into action

Once you have the themes, keep going in the same conversation:

• "Draft a reply to each review under three stars. Keep each under 60 words and don't make promises about refunds."
• "Write a short note for the front-desk team about the top complaint."
• "Which activity gets the most complaints about gear?"

Your assistant drafts the text. You review it and post it yourself, because it can read reviews but can't reply to them for you.

Limits to know

• Counts are approximate. Your assistant reads the records it's given and groups them by judgment. Treat counts as a strong signal, not an audit, and ask it to show the reviews or calls behind a theme.
• Large periods may be sampled. If a range holds thousands of reviews or calls, your assistant may read them in pages. Ask for a shorter window or one activity to get a more thorough read.
• Only recorded calls appear. It can read calls that TripWorks recorded and summarized, such as AI callbot calls and connected phone systems. Calls outside those systems aren't there.
• Guest details are private by permission. If your permission group can't open a call, your assistant can't either.
• Read-only. It can't reply to reviews or call guests back.

Frequently asked questions