Operational AI,
by Nomadia
Service calls orchestrate themselves. Your teams make the calls.
In Field Service Management, AI doesn't transform “a company”: it changes a specific working day. Pick a role: here is what changes, and what runs behind it.
- 4,000 companies
- 200,000 professionals coordinated by AI every day
- Solutions deployed in 92 countries
Which role do you want to transform?
He has just finished his last job. Forty minutes of data entry are still waiting, hands covered in grease, in the van. Or tonight, at home.
One photo on arrival: the false report is ruled outAdd-on
He takes a photo on arrival; the AI compares it with the original photo and rules out the false report. Trip saved, decision logged.
45 min → 2 min per false report
He photographs the fault, the AI tells him what to doAdd-on
One photo: equipment recognized, fault identified, repair recommended with the documentation. No need to call an expert.
25 min → 5 min per fault
His completion photo is proofAdd-on
The completion photo is analyzed: not fixed, not closed. Quality is proven, no longer disputed.
20 min of after-the-fact dispute → 2 min per closure
He dictates instead of typing, and the AI fixes the restIncluded
He dictates, and the transcription fills in the fields of your data model. Typos and abbreviations are normalized: the data becomes comparable.
12 min → 3 min per service report
Van inventory in a few photosAdd-on
A few photos of the stock: items counted, inventory updated, missing consumables flagged before the route starts.
60 min → 5 min per inventory
She opens yesterday's service reports and rewrites them. One by one. Because as they stand, none of them can go to the customer.
The report arrives ready to sendIncluded
Structured, complete, in the right vocabulary. She approves and sends it, instead of retyping.
15 min → 2 min per customer report
The customer gets it the same dayIncluded
Turnaround time is no longer what gives when the workload peaks.
Sent 2 days later → same day
Incoming requests arrive already qualifiedAdd-on
The AI agent qualifies each request within the caller's access rights. She handles cases, not sorting.
5 min → 1 min 30 s per call
An emergency comes in at 10:12 am. He redoes his schedule by hand, moves six service calls, forgets one, and calls two customers back to apologize.
The flood of reports qualifies itselfAdd-on
Photo analyzed, incident confirmed, emergency assigned, technician dispatched by certification and proximity. Manual sorting disappears.
10 min → 40 s from photo to assignment
Only feasible time slots are offeredAdd-on
The agent queries the engine in real time: a slot that would break the route is never offered.
10 min of checking per appointment → 0
Incoming calls are pre-qualifiedAdd-on
The AI agent qualifies calls and answers how-to questions. The call center handles incidents, not training.
5 min → 1 min 30 s per call
Someone's absent? A proposal to approveIncluded
The agent spots the affected service calls and proposes a reallocation that respects certifications. He approves, and the routes are back on track.
60 min → 5 min per absence
Customer calls are handled in one goIncluded
Customer identified, job created, slot compliant with the contract, all in a single call.
12 min → 4 min per call
A key customer calls to complain. He has no immediate visibility: he hangs up, calls someone to go and check, and calls back two hours later.
He asks the question and gets the answer, with figuresIncluded
Overdue jobs and penalty costs: the summary arrives in seconds, on today's data, not yesterday's BI.
45 min → 1 min per summary
The backlog clears, line by lineIncluded
The agent lists the overdue backlog and proposes the order in which to clear it, ready to approve.
90 min → 10 min per backlog review
He acts in the same conversationIncluded
He asks for a reschedule, the engine recalculates, he confirms. The customer is still on the line.
Answer obtained during the customer call
A contract is being signed next week. “We'll need three people per territory, so fifteen.” Nobody around the table knows if that's true.
The contract is costed before it's signedAdd-on
The contract's jobs are fed into the engine, which returns the capacity required, by area and by skill.
3 days → 1 h per contract costing
Annual preventive maintenance: can it be covered, and by how muchAdd-on
Maintenance plan tested against capacity, share reserved for emergencies, growth scenarios simulated on your data.
5 days → 2 h per study · 4 days → 1 h per scenario
Hire or train? Decided on dataAdd-on
Missing skills by territory, volumes to cover: training and hiring are sized, not debated.
3 days → 1 h per study
The call center costs less, at the same service levelAdd-on
A share of requests is absorbed by the AI agent, 24/7. The savings show up in full-time equivalents.
6 min per call → 0 on the absorbed share
A new business form in minutes, not ticketsIncluded
The business team describes the need, the AI generates the form and the report template. IT keeps approval and access rights.
3 days → 30 min per form
He waits for the call center to open, sits through hold music, gets a four-hour window on a Tuesday, and will have to call back if he changes his mind.
He books his appointment whenever he wantsAdd-on
An AI agent converses in natural language, on your channels, at any time.
6 min call + queue → 2 min, 24/7
The agent talks to the engine, not to a calendarAdd-on
The slots offered aren't just free: they fit into the route. The appointment will be kept.
Feasible slots only
Or he reports it with one photoAdd-on
One photo, an instant confirmation. The client receives an already qualified request.
4 min (often abandoned) → 40 s
Five modules, one mechanism underneath
Each one was born from a real customer pain point and co-built with our customers. None of them is a technology showcase.
| Module | What it does | For whom | Pricing | Demo |
|---|---|---|---|---|
| Smart Ticketing | A photo becomes a qualified, prioritized, assigned ticket, and resolution is verified by image. | Back office, technician, end user | Add-on | Video · 1 min |
| Smart Booking | Your customers book appointments 24/7, on slots calculated by the engine: feasible, not just free. | End customer, planner, leadership | Add-on | Video · 1 min |
| Strategic Planning | Feed in the jobs of an upcoming contract and read off the capacity, skills and hiring required. | Operations leadership, HR | Add-on | Video · 1 min |
| Smart Reporting | The technician's voice becomes structured, normalized data that's ready to use downstream. | Technician, back office | Included | Video · 1 min |
| Smart Control | Your managers query activity in natural language, on today's data, and trigger rescheduling. | Manager, planner | Included | Video · 1 min |
Why are some included and others not?
Our rule is simple and we stand by it: anything that improves a task you already do is included: that's the product. Anything that removes a task, or makes a new one possible, is an add-on, because its return is calculated separately. No AI module is billed on usage.
Want to see these modules applied to your own data? Our experts will walk you through a demo.
Request a demoAI does the math. Your teams keep the judgment.
The question that decides adoption. Here's the straight answer.
What goes away
- Redoing the schedule by hand at every disruption
- Sorting reports one by one to qualify them
- Re-entering in the evening what was done during the day
- Rewriting service reports before sending them
- Calling someone to find out where a case stands
- Guessing the resources a contract needs
What stays, and gains value
- Handling situations that fall outside the rules
- Keeping a commitment made to a difficult customer
- Deciding when two emergencies clash
- Judging the quality of a complex service call
- Negotiating a contract knowing your margins
- Upskilling a team
Always the same chain, whatever the module
A signal comes in, the engine decides, the action goes out, the proof comes back. The AI input layer changes. The link that decides never does.
- Signal: photo, voice, question or request, from the field, a manager or a customer.
- Structuring: vision, speech and language; AI prepares the data.
- Decision: the Nomadia optimization engine, 200+ business constraints, deterministic computation, 30 years of R&D.
- Action: assignment and rescheduling, directly in your tools.
- Proof: control and explanation of the choice.
- Humans handle the exceptions and can override any decision.
More than 200 native business constraints, arbitrated in every calculation
Certifications, contractual time windows, vehicle capacities, mandatory breaks, access zones, SLAs. Native: delivered out of the box, not built on request.
The one question that separates two vendors who both claim to do AI.
Calculated, not generated
Your routes are produced by a deterministic engine, fed by predictive models trained on your history. Never by a language model.
MCP connectors at the heart of the engine
Agents don't access your database: they call authorized functions, within the user's access rights. Every figure displayed comes from the calculation.
Your data stays yours
Your field data stays in your Nomadia environment, hosted in the European Union. ISO 27001 certified.
Explainable by design
A rules-based engine can tell you which constraints produced an assignment: exactly what the AI Act will require by December 2027.
Lean AI, true to our mission
Nomadia is a mission-driven company. Performance doesn't justify any footprint: our rule is to use the lightest model that does the job.
The right model for the right task
Routes are calculated by a deterministic engine, using a fraction of the energy of a generative model. Generative AI is only used where it is irreplaceable: voice, vision, language.
A monitored footprint
We track the carbon footprint of the AI tools we use, size models to actual needs and host in the European Union.
Optimization, the biggest carbon lever
Every optimized route means kilometers avoided for our customers. Environmental impact visualization is built into the solution.
AI that reassures your IT team too
No black box, no integration project, no data leaving your environment.
Access rights and scopes respected
Agents call authorized application functions, within each user's access rights. Your existing permissions apply, without exception.
Your data doesn't leave
Environment hosted in the European Union, ISO 27001 certification, training on anonymized data, GDPR compliance.
No integration project
The modules are switched on in your existing environment. And when the official tool is the fastest again, shadow IT fades away on its own.
Auditable, AI Act-ready
Explainable decisions, retained logs, built-in human oversight: the December 2027 requirements are part of our architecture.
Nomadia named in the Gartner® Market Guide for Field Service Management
“Cloud-deployed, comprehensive scheduling capabilities (its scheduling engine embeds over 200 constraints), environmental impact visualizations and a set of multilingual AI recommendations and planning capabilities designed to offer explainability to accelerate adoption.”
What we're asked most often
Everyone says they do AI. How can I check?
Two questions are enough: how many business constraints does the engine arbitrate, native or custom-built? And how long has it been running in production? Nomadia: 200+ native constraints, 4,000 companies, 92 countries.
What is an “AI by design” FSM solution?
It's a solution with AI at its core, not a layer added on top. At most vendors, a chat assistant has been bolted onto an existing scheduling tool. At Nomadia it's the other way around: the optimization engine has been the product from day one, and the new modules plug into it. In practice, the result is calculated and reproducible, not generated.
Twenty years of AI: isn't that outdated technology?
It's twenty years of field data to train the models. The 2026 modules (voice, vision, agents) don't replace the engine: they feed it. Replacing it with a language model would mean losing reproducibility and explainability.
Will my planners be replaced?
No. What goes away: building the schedule by hand and re-entering data. What stays: handling exceptions, customer commitments, steering operations. Planners who lead the rollout become its strongest advocates.
Does the AI respect my existing business constraints?
That's the starting point, not an option. The best way to check is to send us your ten toughest rules, the ones your planners say no tool could ever handle.
Where is my data stored?
In your Nomadia environment, hosted in the European Union for the European market. Datasets used for training are anonymized, in line with GDPR compliance. Nomadia is ISO 27001 certified and favors open-source tools and in-house development.
Where do you stand on the AI Act?
Assigning tasks to workers falls under Annex III of the EU AI Act, with a compliance deadline of December 2027. This requires qualified human oversight, informing employee representatives and retaining logs. An engine built on explicit constraints is structurally better suited to these requirements than a language model.
Do these modules work with my current Nomadia FSM solution?
Yes. They build on the engine and the data you already use. Smart Reporting and Smart Control are included in your license. Smart Ticketing, Smart Booking and Strategic Planning are available as add-ons. No module is billed on usage.
Bring us your worst scheduling case
Your ten most twisted rules, the ones no tool can supposedly handle. The best starting point for a demo.
The figures shown on this page are estimates based on customer feedback. They vary by industry, route density and deployment scope.
