Case study · Field sales
AI agents that check a sales team’s day overnight.
meFolia is the field sales app we designed and built. While the reps sleep, five AI agents read the day’s visits and orders, write up what happened and plan tomorrow. Managers start the morning with answers instead of spreadsheets.
At a glance
Built for field sales teams and the people who manage them.
5
AI agents, each with one job to do overnight
GPS
Check-ins verified inside each store’s geofence
0
Actions taken by the AI without a person saying yes
Web + mobile
One app for reps in the field and managers at the desk
Built to be trusted
The agents suggest. A person decides.
Propose, never act
No agent approves an order, sends a message or edits a customer. A person accepts or rejects every suggestion.
No names leave the app
Reps become “Rep 1”. Names, phone numbers and addresses are never sent to the AI model.
Every run on the record
Each agent run is logged with the model, the time it took and the data it asked for.
Meet the night shift
Five agents. One quiet night. A clear morning.
Order Check
A second pair of eyes on every order. It compares the discount and quantity with the store’s history and checks where the rep was when the order went in.
Next Best Action
One clear next step for every store on tomorrow’s route: a visit, a call or an offer. The rep accepts or rejects each one.
Day Review
Every rep’s day written up for the manager: stores visited, gaps in the day and anything worth a second look.
Follow-up Drafter
No missed follow-up after a visit that ended without an order. It drafts the message, and a person decides whether to send it.
Report Summary
A week or a month of field work turned into a short report in plain words, with the numbers behind it.
Visit confidence score
Every visit gets a score from check-ins inside the store’s geofence and time on site, so the agents start from facts, not from what was typed in.
Product
Inside meFolia.
The journey
From challenge to results.
Challenges
- Managers found out about missed visits days later, if at all.
- Visit reports were typed in, so a real visit looked the same as one logged from the car park.
- Unusual orders, like heavy discounts, slipped through approval.
- Reps had no clear view of which store needed them most tomorrow.
Goals
- Proof that each visit happened at the store.
- A second look at every order before it is approved.
- A short, readable summary of every rep’s day.
- People stay in charge: the AI suggests, a person decides.
Actions
- Built geofenced check-ins measured against each store’s approved location.
- Added five AI agents that read the day’s visits, orders and follow-ups every night.
- Gave the agents read-only access to business data and one table of their own for suggestions.
- Removed names, phone numbers and IDs before any data reaches the model.
- Logged every agent run with the model, time taken and data requested.
Results
- Managers start the day with a written review of every rep’s day.
- Unusual orders are flagged to a supervisor before approval.
- Each rep gets one ranked next step per store, to accept or reject.
- Field work, CRM, invoices, helpdesk and team chat in one app.
The platform
Every business tool, one app.
Field Work
CRM
Quotations & Invoices
Appointments
Helpdesk
People & Org Chart
Chat & Calls
Calendar & Events
Communities
Let’s talk
Want AI agents working on your field data?
Share a few details about your project and our team will get back to you with next steps.
What happens next
- We reply within one business dayA senior engineer reads every message, not a sales bot.
- 30-minute discovery callWe map your goals, data and constraints together.
- A clear plan and estimateScope, architecture and timeline, usually within a week.
