FlowsBook a callSign inCreate free accountSign up
AI employees and automation

An AI employee for operations: what it watches, plans and reports

What an AI employee for operations does for depots, field crews and site teams: road and weather warnings, daily routes, fair territories and check-ins.

Part of the guide AI employees

By the xMap team · · 4 min read

An ai employee for operations does the watching and planning an operations lead would do first thing each day if they had the time: it checks the roads and the weather around every site, plans the order of the day's visits, shares work fairly between crews, and tells the right person only when something needs a decision. It works before the shift starts, every day, and stays quiet when nothing has changed.

This article is for operations managers in delivery, field service, facilities, merchandising and logistics. It describes four jobs operations AI employees do well, a made-up morning at a service company, what to be careful about with live data, and the blocks in xMap Flows that build each job. For the wider picture, see the AI employees for field operations guide.

Four jobs that suit operations

Watch the conditions around every site

Operations problems often start outside the company: a road closed by works, an accident near a depot, a storm due on a day of outdoor jobs. An AI employee checks each site's surroundings on a timetable and sends one short list of what affects whom. It can read road closures and incidents, live road speeds with the jam factor (how congested a road is from 0 to 10), and the weather forecast up to 16 days ahead.

Plan the day's visits

Given the day's stops, it puts them in a sensible order from a start point, by car, on foot or by bike, and adds a directions link for each part of the route. With several reps, it can plan one route each.

Share work fairly

When a list of accounts, outlets or jobs comes in, it splits them between a number of reps by area into compact territories of about equal size, and names each one by direction from the middle.

Ask before acting

Some decisions need a person: reschedule a job, send a crew elsewhere, approve overtime. An operations AI employee can stop and ask someone by email, WhatsApp or Slack, wait for the answer, and carry on with it.

A worked morning

This example is made up. A facilities company in Singapore sends technicians to about forty office buildings a day across the island. The dispatcher used to spend the first hour checking the news, the weather and a map before calling technicians.

Now an AI employee runs at 05:45. It takes the day's jobs from the dispatch sheet, checks road incidents and live speeds near each building, and reads the forecast for heavy rain. It splits the jobs between the six technicians on shift by area, plans each route, and sends each technician their list with directions. It sends the dispatcher one message: two buildings have a road closure next to them, and rain is expected after 15:00 in the west. The dispatcher moves one roof job to the morning and the day starts.

Through the day, technicians check in at each stop on their phones, noting whether the visit was done and adding a photo. The dispatcher sees the day's check-ins stop by stop, and the AI employee's note in the Team inbox the next morning says what its run found and sent.

Be honest about what the data says

  • Live speeds are not car counts. Traffic Flow gives the current speed, the usual free-flow speed and the jam factor. It does not count vehicles. Hourly car counts going back to 2021 exist only for roads in the United Arab Emirates, through Regional data.
  • Checks happen on a timetable. A closure is seen at the next check, not the second it happens. Choose how often to check by how fast you would act.
  • Routes are estimates. A planned route uses roads and typical speeds; a closure discovered on the way still needs a person.
  • Footfall arrives late. If you use how busy an area is to plan staffing, remember footfall is people-hours of presence and arrives about ten days after the day it measures.

Mistakes to avoid

  • Alerting on everything. A list of every incident in a city is noise. Filter to what is within a few hundred metres of your sites.
  • One message for everyone. Each technician needs their own stops; the dispatcher needs the exceptions. Split the output.
  • Running it too late. A road warning at 09:00 is history. Schedule it before the shift and in the time zone of the sites.
  • No owner. Name one person who reads its notes and acts on its questions, or the warnings go nowhere.

How to build it in xMap Flows

The Singapore morning above is one flow of blocks:

  1. Schedule: every weekday at 05:45, Asia/Singapore time.
  2. New rows or My data: the day's jobs, with an address or position each.
  3. Traffic incidents around each job, and Traffic Flow for live speeds.
  4. Weather forecast for each job: rain expected today and when.
  5. Split by territory between the technicians on shift.
  6. Route planner: one route for each technician.
  7. Email or WhatsApp: each technician's stops; a separate short message to the dispatcher with the exceptions.

Add Ask a person where a decision is needed, such as whether to move an outdoor job when rain is forecast. Templates to start from: Road closures in my city, Merchandiser visit plan for the week and Split my city into fair sales territories.

Questions

Can it re-plan routes during the day?

It plans when it runs. You can run it again, or set it to check every 15 or 30 minutes during working hours, but it does not follow vehicles live.

Can technicians reply to it?

They can check in at each stop on their phone. For questions, the AI employee can ask a person by email, WhatsApp or Slack and use the answer.

Does it work outside cities?

Roads, weather and places are worldwide, but coverage of incidents and live speeds is thinner on minor rural roads.

Who sees the routes?

Whoever you send them to, plus members of your organization who can see the flow and its results.

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.