Start collecting data now!

AI needs training data. Even the most sophisticated AI model is unable to deliver results from nothing. If companies had started collecting and processing data at an earlier stage, they would now have a head start with classic or generative AI. Since many companies have not done this, they now do not have a data basis – and are putting corresponding projects on hold. Unfortunately, they then often postpone the decision as to which data should be collected and how it should be processed. The consequences are disastrous.

Technologies like GenAI are currently changing a large number of markets from the ground up – just as the Internet, mobile business, and the cloud have done in the past. While these have also had a lasting impact on the economy and made many companies successful, they have also caused difficulties for many an established company. Just as we are now looking back at the early days of these technologies, in a few years’ time we will be reflecting on the current phase of AI and GenAI. This is why companies need to set the course for a successful future in this data-driven world and start investing in data now.

However, there is a key difference between investing in a company’s own data management and other investment decisions: a company that acquires new machinery or opens a new subsidiary is usually able to calculate the return on investment based on past experience. However, companies that start by collecting, cleansing, aggregating, and supplying data can often only guess at its potential – its real value only becomes apparent later. It is only when the data are available and analysed that their true potential is revealed. For example, take the question of whether telephone data from a call centre can actually improve service staff performance. Or whether the tracking data for a landing page really can yield ideas for optimising a marketing campaign.

And yet the advice I give to companies is straightforward: be courageous and – with due regard to all data protection regulations and regulatory requirements – start collecting, collecting, collecting straight away. And by this, I mean data and experience. At the same time, companies need to put the right conditions in place to ensure an optimal flow of data within the organisation, otherwise bottlenecks in the provision of data will put the brakes on AI projects further down the line.

And the best response to those who object that we still do not know what all this data will be used for is: “but I do know what will happen if we do not start collecting data now.”