NEW RESEARCH | Five keys to successful innovation networks

Developing digital innovations often requires collaboration between multiple actors. But collaboration is no easy feat. So how do successful innovation networks work?

An AI system to make shipping safer and more sustainable. That was the goal of one of the innovation networks that Adrian Bumann, Chalmers, has studied in his doctoral thesis.

– It is becoming increasingly common for actors from different sectors to work together in networks when developing new digital technologies. Not least because of all the societal challenges we face, which in many cases we can solve with advanced technology. But networking is demanding and not always easy, he says.

The maritime navigation project studied involved actors from very different organizations and companies with expertise in navigation, computer science and telecommunications. The task was to jointly develop algorithms that can analyze ship and GPS data to better predict a ship’s movement and support the navigators on board. To provide the network with a common frame of reference, and facilitate discussions and decision-making, a simulation software was created to visualize AI capabilities.

– It helped participants understand the functionality of the algorithm more clearly, whether they were computer scientists or engineers, explains Adrian Bumann.

Multiple uses

Essentially, people in innovation networks work in the same way as in any collaborative project. They brainstorm together, sketch out what they want to do and work on it – and maybe change their plans along the way, to finally have a result. But there is a crucial difference between building a car and developing AI software, according to Adrian Bumann.

– There are more nuances and more flexibility when working with digital innovation. Data and algorithms developed for a specific purpose can often be reused for other purposes.

– So, during the course of the project, new uses can be discovered and targets adjusted in real time.

Difficult with hierarchical governance

When several organizations with the same knowledge base are working together, it usually works well to have one person coordinating the group and the work. But in an innovation network, where everyone comes from different sectors and has different types of knowledge, no one can know everything.

– Knowledge – and therefore control – is then dispersed across the network. So in an innovation process where new advanced technologies are to be developed, coordination is instead bottom-up, says Adrian Bumann.

Was there anything in your research that surprised you?

– Yes, it is very challenging to find good data. That is, data that is reliable, accurate and available for the purpose you need it for. The most complicated AI algorithm is not useful if you don’t have enough qualitative data to feed it with, says Adrian Bumann.

Contact adrian.bumann@chalmers.se