Most digitalisation and AI projects fail not because of the technology, but because of the factors surrounding it. There is no owner, no measure of success and no plan for how to roll the solution out into regular operation. Daniel Pecina from Trask, which supplies IT solutions to Škoda Auto, explains why it’s best to start with a single process that’s causing problems, what a university can offer a large company, and why a pilot project shouldn’t just be a ‘trial run’. He will be speaking on this topic at the Technical University of Liberec (TUL) as part of the University Doctoral School series.
Where do you see the greatest scope for collaboration between industry and universities and research organisations?
Mainly where a company is tackling a problem for which there is as yet no ready-made solution on the market. Typically, this involves analysing specific industrial data, optimising complex processes, or validating new technologies before the company invests in them on a large scale. I also see great potential in joint student and PhD projects. The company provides a real-world brief and data, whilst the university contributes the methodology and the time for in-depth investigation. We also greatly value our collaboration with your technology institute on Proof of Concept (PoC) projects – that is, verifying a specific technological process using real-world data. This gives our company time to assess scalability, operational viability and support options under real-world conditions.
In your view, what can a research environment offer a large company that the company would find difficult to create on its own?
Time and space to think deeply, without the pressure of quarterly results. Companies are often focused on delivery and operations, so they lack the capacity for systematic experimentation, testing hypotheses and keeping up with the latest research. A university brings an independent perspective, access to experts from various disciplines and, last but not least, talented individuals whom the company can get to know even before they enter the workforce. At the same time, it provides access to innovative technologies, which are difficult for a company of our type to secure until we have a commercial use for them.
How should a company proceed if it wants to start with digitalisation, AI or process modernisation, but doesn’t know where to begin?
Don’t start with the technology, but with the problem. Choose one specific process that is currently causing issues, is measurable and has a clear owner, and use it to demonstrate the benefits. At the same time, it’s a good idea to look at the state of the data right from the start, and consider who will use and maintain the solution once it’s up and running. A small, well-chosen first step will teach you more than a grand strategy on paper. This is a tried-and-tested approach that ensures a company embarking on technological innovation does not immediately take on too much in the form of change management – that is, communicating and implementing the changes that a new approach to digitalisation will require.
What is the most common factor in determining whether a digitalisation or AI project within a company will succeed or end up as nothing more than a pilot?
It is usually not the technology itself that is decisive, but what surrounds it. Pilots are often carried out ‘on a trial basis’, without a clear business owner, without pre-defined success metrics, and without a plan for how to integrate the solution into existing systems and processes. Successful projects have the entire accompanying change management process clearly defined from the outset – that is, support from senior management and the involvement of the people who will be working with the outcome. They take operations, security and data quality just as seriously as the operational model itself. A pilot should be the first phase of deployment, not a standalone experiment that only begins to address the above issues once it has concluded. With AI in particular, the implementation steps in software development are very rapid. It is therefore advisable to have the accompanying changes and steps carefully planned and secured, so that there is no delay between the launch of a digitalisation project and its actual use.