Insights · Perspective
Perspective · · 7 min read

What is Mobility as a Service (MaaS)? Origin, meaning and where AI fits

Mobility as a Service bundles public transport, taxis, bikes and scooters into one app to plan, book and pay. Where the term came from, what it means, real examples, and the integration and AI work behind it.

The short answer

Mobility as a Service (MaaS) means you buy mobility, not a car or a single ticket. One app lets you plan a trip, book it and pay for it across every transport option a city has: metro, bus, train, taxi, ride-hailing, car sharing, bikes and e-scooters. Some MaaS services also sell monthly bundles, like a phone plan for getting around.

The idea is simple. The private car is convenient because it is always there and needs no planning. MaaS tries to give the same convenience by joining many services into one experience, so living without a car becomes easy.

Is it a well-known term?

Yes, in its own world. Transport authorities, city planners, operators and the smart-city industry use the term every day. The European Union, national ministries and large operators run MaaS programmes, and there is an industry body, the MaaS Alliance. For the general public it is still a niche term: people use MaaS apps without knowing the name.

Where the term came from

MaaS was born in Finland around 2014. The Finnish engineer Sampo Hietanen, then at ITS Finland, is widely called the father of the concept, and a master’s thesis at Aalto University in Helsinki the same year described it as a proposal for the city’s public administration. Sweden had already tested the idea in Gothenburg with the UbiGo pilot in 2013–2014.

In 2015 Hietanen founded MaaS Global, and in 2017 it launched Whim in Helsinki, one of the first apps to sell transport bundles as a monthly subscription. Finland then changed its law so that transport operators had to open their data and ticket sales through APIs. That law mattered more than any app: without access to the operators, there is no MaaS.

A useful model: the five levels of MaaS

Researchers at Chalmers University in Sweden describe MaaS in levels of integration. It is the clearest way to judge what an app really does:

  • Level 0, no integration. Separate apps for each operator.
  • Level 1, information. One journey planner across modes, with schedules and prices.
  • Level 2, booking and payment. Find, book and pay for the whole trip in one place.
  • Level 3, the service offer. Bundles and subscriptions across operators.
  • Level 4, societal goals. The city uses MaaS to steer behaviour, for example pricing that rewards public transport or reduces congestion.

Most apps in the market today sit at level 1 or 2. Very few reach level 3 in a sustainable way.

Examples

  • Whim, Helsinki. The pioneer. It proved the idea and attracted large investors from the car and energy industries, but subscription bundles were expensive to run, and MaaS Global filed for bankruptcy in 2024.
  • Jelbi, Berlin. Run by the public transport operator BVG. It joins public transport with sharing services and has physical mobility hubs at stations.
  • WienMobil, Vienna. The city’s transport company combines public transport, bikes, car sharing and taxis in one app.
  • Journey planners from Google and Apple, and super-apps like Uber. They reach level 1 or 2 for many users, which shows the demand, and also the pressure on public operators.
The lesson from the first decade: MaaS works best when the city or the public transport operator sits at its centre, not as a stand-alone start-up selling subscriptions.

MaaS is an integration problem first

Behind the simple app is hard plumbing. Every operator has its own systems for schedules, live positions, availability, tickets and payments. A MaaS platform has to connect all of them, and keep working when one of them changes. In practice that means:

  • Open data standards. GTFS and GTFS-Realtime for public transport, GBFS for bikes and scooters, NeTEx and SIRI in Europe.
  • Booking and ticketing APIs. Standards such as TOMP-API, plus many custom APIs per operator.
  • One payment and settlement layer. The user pays once; the platform has to split revenue fairly between operators and reconcile it every day.
  • Identity and consent. One account, one profile, discounts and entitlements that every operator trusts, under data-protection rules.
  • A canonical data model. Trips, vehicles, stops, prices and users described the same way across every source, so the rest of the platform does not care where data came from.

This is classic API-led and event-driven integration. Most failed MaaS projects did not fail on the app; they failed on the cost of connecting and running everything behind it.

Where AI fits

AI does not replace this foundation. It sits on top of it, and only works if the data underneath is clean and connected. Where it adds real value:

  • Demand prediction. Forecast where people will travel by hour and place, so operators place vehicles and bikes before demand appears.
  • Smarter trip planning. Routes that learn from real delays, crowding and the user’s own habits, not only from the timetable.
  • Dynamic offers. Bundles and prices that reflect real use and city goals, the step that makes level 3 and 4 affordable.
  • Operations. Fleet health, maintenance planning and incident handling across many operators. This is the daily work of fleet and workforce operations on Cooriroo, the AI Operations Platform we build.
  • AI agents and assistants. A traveller asks in plain language, “get me to the airport by 9, cheapest option”, and an agent plans, books and pays through the same APIs. Protocols such as MCP make those APIs usable by assistants, with the same identity and permission rules as any other client.

What this means for cities and operators

MaaS is not a single app you buy. It is a platform decision: open data, shared APIs, one payment layer and clear governance on who may expose what to whom. Build that layer well, and the app, the bundles and the AI on top become much cheaper to add and change. Skip it, and every new operator or feature becomes a new project.

At devtech.pro we have spent 25 years on integration, and today we build the AI layer on top of it. If you are planning a mobility platform, a city data model or AI on top of transport data, we are happy to talk.

Related reading

By Bassem Fawaz · #mobility #maas #smart-city #integration #ai-agents

Building a mobility platform?

We connect operators, payments and data into one platform, and add AI where it saves real time and money.

Talk to an expert