Why so many AI projects fail and how you can do better

Digital transformation has long been the biggest challenge for companies and organizations worldwide. Many entrepreneurs are visibly struggling with it and are dissatisfied with the results they have achieved so far. How can so many digital projects fail and companies be so dissatisfied with what they have achieved? 

The answer is based on 2 aspects: Knowledge and expectation.

There is often a lack of knowledge about what a (digital) transformation really is and how it needs to be planned, orchestrated, and implemented to be “successful”. A lack of knowledge then leads to false expectations of the digital transformation. What does “successful” actually mean? If you don't know what transformation is, how it works, and what success means for you, realistic expectations are impossible. Disappointment is inevitable.

The error is at the beginning

Most “digital strategies” in companies and organizations start with a “digital pet project” (usually from the boss). Be it an app that is desired now “because everyone has one” or a digital showroom that is supposed to turn a dusty textile company into a supposedly dynamic brand. In both cases, there is a “why” but with a useless justification.

If you compare the above examples with soccer, then in both cases it’s like starting the game in the 85th minute and then wanting to see the decisive goal for your team, which is scored in the 90th minute.

A lot has already happened before the 85th minute

Up until the 85th minute, however, there was plenty of play, the teams worked hard, tried a lot, and played many a misplaced pass and many a dream combination. Sometimes they drove the crowd and the coaches to despair, sometimes to enthusiasm. And not to forget: The game was preceded by many weeks and months of training, preparation matches, trips to training camps, and so on.

You see, before the 85th minute, even before the game started, there had been enormous efforts. And even if the team then wins the game in the 90th minute, that is only the last minute before the 1st minute of the next game.

It all comes down to long-term planning

It's the same with digital transformation. Not only do you need to know the rules of the game, but you also need to plan for the long term instead of just focusing on the last 5 minutes. You need the right coach and plenty of teammates who are all trained in their respective positions, familiar with the tactics, and equipped with match practice. And above all, they all need one thing: the confidence that they can do it together.

It’s necessary to see digital transformation as a very long and complex process. It’s also a process that never ends (“after the game is before the game”), that needs to be precisely planned (strategy roadmap) and that requires a lot of tactical study and analysis (data). Digital transformation requires experimentation, trial and error, failure, and new beginnings (innovative, agile organizational culture). If I don't understand all this but expect victories with a difference of 4 or 5 goals, then my disappointment about the 1:0 victory in the 90th minute (tools and applications) will be huge. 

That is precisely the problem: the digital tools are “only” the goal in the 90th minute.   

If I have neither an idea of the future (vision) nor a plan of how I can achieve this future (mission), then I cannot determine how I will get there (strategic roadmap). But then every expected result (whether a 1:0 win or 6:2) is like playing the lottery.

Tools alone do not make a transformation

So if I start with a digital tool as a goal (for example, wanting to have an app or a digital showroom), this is not a digital transformation, but a lottery. It may go well, but probably not. While in soccer many people think they know exactly how it should work no one thinks that way about digital transformation. A few people venture forward and demand the goal from the 90th minute, and they want it now.

Over 80% of companies state that they find it difficult or very difficult to develop digital products and services or that they do not develop any at all.

According to a Bitcom study, 52% of all companies in Germany state that they are not even considering the use of AI – simply because they do not have an AI strategy. On the one hand, this is very sensible – at least they won't be spending their money on useless things – but on the other hand, they run the risk of gambling away their future. In soccer, I would say: If I don't deal with training, analytics, game organization, and match plan, relegation is inevitable.

ChatGPT yes, but no AI strategy

Companies should seriously ask themselves why they have marketing and sales strategies, IT strategies, or even strategies for logistics – but no digitalization or AI strategy. This seems all the more strange given that most companies have identified digitalization and AI as the key technologies of the future. The fact that some have at least started to think about the topic is shown by the development of ChatGPT usage in companies. This has risen significantly from 150‘000 in January 2024 to 600‘000 in April 2024. Applied to the soccer example: Some people have at least subscribed to “Kicker” and can read the table.

Set up your match plan

Back to the initial question: Why are so many companies dissatisfied with implemented digital tools and solutions? Well, if these digital / AI projects were piloted in the companies but not rolled out throughout the entire company (because there was no match plan) and thus only had no chance to exploit their full potential as an “isolated solution” (the boss's favorite project), real success could only fail to materialize.

All instead of a few

It was then just the “boss project” and not the “project of all”. There are also other internal company reasons: an often fragmented and sometimes outdated IT infrastructure, a lack of interfaces between the systems, and insufficient quality of incomplete data. This makes it all the more important to think about digitalization and AI not for a few but for everyone.

In soccer, the club board is not the only one on the pitch with a coach. A team needs everyone – club, coaching team, analysts, tactics experts, doctors, nutritionists, players, including those on the substitutes' bench, even fans and spectators – to have the same goal (vision) in mind. If everyone believes in this goal, if everyone accepts the match plan (mission & roadmap), then the 1:0 in the 90th minute will be a success and celebrated accordingly. Then the team also knows that it will score its decisive goal in the 90th minute – no matter how many times it has missed the goal beforehand (mindset). The match plan is clear, the tactics are adjusted on the go – after all, you're not alone on the pitch.

Digital transformation must become the central responsibility of company management, with the full backing of the owners (because it will take a long time until “the 1:0 falls” and will require a lot of resources). Together with a broad range of stakeholders, the management develops a (digital AI) strategy roadmap and ensures that everyone understands this roadmap and believes in the goal. They also ensure that all employees know their role and that they must and will score the goal in the end.

Knowing the data is essential

How the goal is scored is determined by the game – but that also means that although we start the game with the firm belief that this goal will be scored, we don't know at the start whether we will score it ourselves or whether it will be an own goal, whether it will be a long-range shot, a header, a clearance or a penalty. But everyone knows that it will be 1 to 0. In addition to the IT infrastructure, the existence and nature of the data are essential. How can I analyze my game and that of my opponent if I know neither myself nor my opponent? AI consists of algorithms, and these cannot work without the corresponding data. The AI strategy therefore includes determining which data the company systematically collects, where it is stored, how we obtain information from it, who has access to it, and how it can be secured.

Data quality must also be ensured, as databases are often full of duplicates, abbreviations, missing values, or different units of measurement. A lack of data quality ultimately means that AI applications either don’t develop their full potential or, in the worst case, even deliver incorrect, disappointing results. The example shows that: In the end, even a last-minute win can be disappointing.

Conclusion: Involve everyone and start before the 85th minute

Once again, the appeal goes out to entrepreneurs and business leaders: Rely on the intelligence of the masses. You’re not just playing your game for yourself, but for many others – especially those who cheer you on in the stadium and buy tickets, jerseys, and live digital broadcasts. Include these “many others” in your considerations and decisions.

Technology doesn’t make people superfluous. Gear your technology even more towards people and use everyone's knowledge to further expand your technological lead. There are successful models to help you do this.

A soccer match does not start in the 85th minute. So don't let your AI strategy start in the 85th minute either.