Executives from IBM, Telefónica, Aggity, Nealis, Libelium and ITI met with the head of Harvard University’s Data Systems and AI Lab at La Harinera in València to analyse how data-sharing ecosystems can create new business opportunities and improve the efficiency of companies and public administrations.
The meeting took place at the event Data Spaces: the value of shared data, organised by València City Council’s Centre of Excellence and Data Office and EDINT together with Smart Business Lab, with Nealis Tech as main sponsor.
The data economy is advancing rapidly, but it still lives with a paradox: much of the information generated by companies and public administrations is never put to use.
“Around 55% of the data produced is ‘dark data’, that is, information that is captured and stored but never processed or used,” noted Alejandro Martínez, project manager at ITI, at the start of the round table, which featured technologists from the Data and AI units of IBM, Telefónica, Aggity, Nealis and Libelium.
In the view of Vidal Martínez, Senior Business Development Manager | Data & AI & IoT at Telefónica Tech, developing data spaces only makes sense if it responds to specific needs. “The first thing we must bear in mind is data governance. Data spaces add a great deal of value if we start from clear use cases that act as the engine of the ecosystem.”
Many organisations already have infrastructures capable of processing large volumes of information, but still struggle to access data or prepare it for specific projects. “Data infrastructures are getting faster and faster at generating competitive advantages, but the difficulty lies in how to access the data and prepare it so that it really serves the use cases we set ourselves,” explained Nuno Maximiano, Director of Select Growth Spain – Data & AI at IBM.
The challenge of breaking down internal silos
Beyond technology, the debate focused on the organisational changes the data economy demands. Many companies still manage information in a fragmented way.
“Sharing that information and governing it jointly is complex, but it is precisely what makes it possible to develop projects with impact,” said José Mª Salido, Data & AI Senior Consultant at Aggity. “When you bring data together within the same ecosystem, you can identify process leaks, inefficiencies or production problems that would otherwise go unnoticed.”
An underused resource
For Marcos Rozas, General Manager of Nealis Tech, the current situation reflects a clear underuse of the data available.
“We are collecting an extremely valuable resource and leaving it parked. Data is the fuel that allows us to carry out predictive analyses capable of saving a client a lot of money.”
Rozas used an analogy to explain the problem: “If we set up a company in which 80% of employees don’t come to work, it clearly won’t go very far. We are doing something similar with data. It isn’t being used as it should be.”
Antonio Jara, CSO of Libelium, argued that the process must always start with the right question: “Once we know what challenge we want to solve, we must look for the right data source.” Combining different sources, he noted, usually translates into efficiency gains thanks to business intelligence.
Trust, key to the future
The debate also addressed trust in a context where artificial intelligence is starting to make decisions autonomously.
“We are increasingly creating more automated services, and a key question arises: who is responsible when an AI agent makes a wrong decision?” asked José Mª Salido.
Antonio Jara raised another issue: the possibility of developing mechanisms that directly compensate those who suffer harm. “It would be very simple if there were a data insurer guaranteeing that if, for example, there is a security breach in decision-making, the companies or people affected would be properly compensated. This would build greater trust in the ecosystem.”
Nuno Maximiano warned that the challenge grows in complex systems based on artificial intelligence. “AI governance involves complex challenges: the data, the models, the agents that take actions… we are still at an early stage and responsibility is not always clear.”
Given this scenario, he called for stronger technological and human oversight: “All systems that make decisions must be monitored very closely. If we don’t get the technological and governance foundations right, the problem multiplies.”
Towards a new relationship with data
The conversation ended with a look at the future of these ecosystems and their impact on decision-making.
According to Marcos Rozas, “we are heading towards a scenario in which we will talk to our data to make decisions and simplify processes that are much more complex today.”
He added: “If agentic AI is increasingly going to act on critical processes, let’s make sure we give it quality data to train it so that it works rigorously. That is why, given where we are heading, sharing data across sectors is becoming essential.”