FlowsBook a callSign inCreate free accountSign up
AI employees and automation

An AI employee that writes your weekly competitor report

A competitor report nobody has to compile: what opened, closed, changed rating or started hiring near you this week, written up and in the inbox every Monday morning.

Part of the guide AI employees

By the xMap team · · 5 min read

An AI employee can write a weekly competitor report on its own: each week it looks at your rivals around your locations, keeps only what changed since last week (new branches, closures, ratings that moved, roles they started hiring for), has xMap AI word the findings using only the figures it found, and emails a short PDF to the people who need it on Monday morning. Nobody compiles anything, and nothing in the report is invented, because every line comes from the week's own results.

This article is for owners, marketing and operations leads who want to keep an eye on competitors without a weekly afternoon of searching. It covers what belongs in such a report, two ways to build the employee, a made-up example for a small pharmacy group, and how to make sure the report stays honest and readable.

What a useful competitor report contains

A competitor report earns its place in the inbox when it is short and only about change. Five sections cover most needs:

  1. Openings. Rival branches found for the first time near your own places.
  2. Closures. Rival branches that are now marked closed.
  3. Ratings that moved. A rival whose rating rose or fell by a set amount since last week, which often follows a change of staff or ownership.
  4. Hiring. Rivals posting jobs near you, especially store managers or opening teams, a sign of growth to come.
  5. What to watch. Two or three sentences on what it adds up to, written only from the above.

Leave out what did not change. A list of every rival with its rating each week is a directory, not a report, and people stop reading it by the third Monday.

Two ways to build it

One employee that does everything

  1. Schedule: every Monday at 06:30.
  2. My data: your own locations.
  3. Nearby places: rivals within, say, 2 km of each location, with Only this brand set to the names you watch, or every place of your kind if you watch the whole market.
  4. What changed: keeps only places whose details changed since last week: closed, a rating that moved by the amount you set, a new phone or website. A second branch through Only new catches places seen for the first time.
  5. xMap LLM: working on the whole list, writes the five sections, told plainly never to add a number that is not in the items.
  6. Make a report: a short PDF with your logo, emailed to the team.

A team of watchers and an analyst

If you already run separate AI employees (one for rival openings, one for hiring, one for road closures near your sites), the Team findings block reads everything they found this week against last week and turns it into findings with their figures: openings and closures by city, a rival's new branches within 1 km of your own places, who is hiring. An xMap LLM step words them and Make a report sends them. This is the Market analyst: it never searches, it only reads what your other employees found, so each watcher can be changed without touching the report.

A made-up example: six pharmacies in Toronto

Toronto's pharmacy market is a mix of large chains and independents. xMap's places database lists 519 pharmacies within 10 km of the centre (November 2026); of the 438 nearest the centre, 119 belong to a chain, the largest with 57 of them, and 319 are independent. That market context is real; the group and its week below are invented.

An independent group of six pharmacies sets up the employee with the three chain brands it competes with most and every independent pharmacy within 2 km of its stores. On a Monday in spring the PDF reads, in short: one new branch of a chain found 700 m from the group's east-end store; one independent near the west-end store now marked closed; a rival two streets from the main store fell from 4.4 to 4.0; a chain posting for a pharmacy manager in the north of the city. What to watch: the east-end store faces a new chain branch a short walk away; the closure near the west end leaves its customers looking for a new pharmacy. The owner asks the east-end manager to check opening hours and the west-end manager to put a welcome offer in the window.

Keeping the report honest

  • Figures come from the items, not the AI. Tell xMap LLM, in its instructions, to use only the numbers in front of it and to say so when a section is empty. Team findings goes further: every figure is counted from your results, and the AI only words them.
  • Say what is missing. Opening dates are absent for some places, so a branch "found for the first time" may have opened a little earlier. The report should say found, not opened, unless the date is there.
  • The first week is a baseline. On its first run the employee learns what exists. The report is about change from the second Monday.
  • Decide how much it does alone. A report that only goes to your own team can be sent without asking. If it goes to a client or a board, set the employee to ask you first and approve each week's PDF.

Reading it well

Give the report ten minutes on Monday and one action per section at most. Over a quarter, the PDFs become a record of how your market moved, which is useful in planning meetings and when you choose your next location. If a section is empty for weeks, drop it; if people ask for something each week, add it. The A weekly brief from your AI employees template is the Market analyst ready to run, Weekly competitive intelligence for my cafes is a single employee for cafes, and Rival openings and closures every month is a monthly version for retail networks.

Questions

Will the AI make up findings if nothing happened?

Not if it is told plainly to use only the items in front of it and to say when a section is empty. With Team findings, every figure is counted from your results before the AI sees it.

Can it tell me why a rival's rating fell?

It can tell you the rating moved and by how much. The reason is not in the data; read the rival's recent reviews yourself if it matters.

Who should receive the report?

Whoever acts on it: usually the owner or regional manager, and the managers of the locations it mentions. A report that goes to everyone tends to be read by no one.

Can I get it in Slack instead of a PDF?

Yes. Swap Make a report for a Slack message, or send both; the findings are the same.

Templates that do this

Read next

More on ai employees and automation · Every article · Templates

Put it to work

An AI employee in xMap Flows does this on live data, on its own timetable. Start from a template or describe it in plain words.