How we create innovative AI solutions at ProSiebenSat.1
by Fahim Alefi, VP Data Strategy
Embarking on the next best AI initiative requires a thorough examination of options — in terms of opportunities but also in terms of feasibility and most important real life, hands-on, money-in-the-pocket value add.
At ProSiebenSat.1, we have developed a process accompanying our teams from idea phase to operation of AI tools or AI driven processes.
We have a few different ways of doing what we call ‘ideation’– coming up with new ideas. One is industry or cross-industry benchmarking, so looking around at what others do. As second is technology scouting: What is happening in the AI area and what new technologies are available. Then the third is the classical way of going through the different departments and talking with them about the possibilities AI gives us. Since AI is such a buzzword nowadays, it makes sense to explain the emphasis we put on it when talking about AI. We help our businesses to automate processes or support decision making based on large amounts of data and more than trivial algorithms.
Transparency is everything — from scoring to decision making. Not only for us but rather for the overall engagement of all involved parties.
We set up a process called ‘Value Creation Process’. It entails scoring mechanism to map ideas and projects and to eventually take them through different stages. A project must reach a certain score to step up the next stage, and as soon as money or resources are involved, a committee or board discusses if we as a company or as a subgroup are going to invest. No matter if it’s ideas from tech experts, business cases or benchmarking, it all undergoes the same process with the same scoring so that we have a transparent comparison of all ideas and projects.
The first thing we look at is the value created for us as a Group, and the importance of the related business area. Is it a core business or a supporting function like HR or finance? Then the second thing is how feasible it is for us. Do we have the right skills? Can we afford the skills if we must look externally?
Eventually, we check for the data available at hand and data to be obtained externally. And this is potentially a huge difference in how data strategy is thought by others. You can often read that organizations start to collect and catalogize data and start to think of use cases within the borders of available data. This might be one angle to look at it, but it definitely shouldn’t be the only way. In fact — and like the standard creativity processes — one should start out of the box with a scenario or a goal you want to achieve, independent from your current position, and continue with thinking how to get there.
Feedback loops and overall success rate
Reviewing and checking whether you are on the right track is key to all sorts of projects. We are often forced to do that because our first attempts usually don’t work out. Therefore, our ‘Value Creation Process’ has a few built-in feedback loops where we revise and change things to make them work.
Regarding our overall success rate, we are as good as predictions imply. As any larger company, we have processes and criteria for investment decisions and we try to set up a business case and calculate some financial KPIs, but that’s all paperwork. As a matter of fact, you cannot look into the crystal ball. Especially for data projects, it’s more trial and error than a classical IT project, in which you know you’re going to substitute one tool with another or create a new process. Very often it depends on the data you have available and your skills to apply the technology or algorithms. The prediction of the outcome is even less reliable than for any other project, at least in our understanding.
Conflict management and killing projects
In case of a major disagreement, we would need to discuss and find a solution to follow within the committees established. But since everything is also new for us and brought to life just last year, for the time being, it’s quite clear what to do and what has the highest priority. Therefore, the projects we are talking about right now are clear and obvious. We were not yet in the situation to delete an item from the list. It’s rather you need to try and find the way to make it work. But we can see already now that scraping an idea or even a project from the list is maybe one of the toughest tasks.
Change management
My personal career was mainly in consulting– seeing things from the outside, only looking at the objective or measurable facts and how things should be. But from the inside, you see that having people want to join an endeavor is one of the most significant challenges. You can force people, but it’s not valuable to have people on board who don’t believe in the project or who do not want to do it. Change management is probably half of the work of bringing something to life.
Based on the interview with EnterpriseAI Consulting.
