AI vs. budget: what we learned debating the real profitability of AI with PwC and Multiverse Computing

Madrid. Almost every company has tried AI by now. Very few have managed to get it into production and see it show up in the bottom line. That was the conversation we brought to Torre PwC on Paseo de la Castellana last Friday, at a roundtable titled ‘AI vs. Budget, the New Challenge for Companies’, organised by El Confidencial and PwC.

Our CEO and founder, Francisco Moreno Balboa, shared the stage with Armando Martínez Polo, Partner at PwC, and Álvaro Alcalde López, Strategic Partner Manager at Multiverse Computing. Three different perspectives (consulting, quantum computing and AI integration) and a surprisingly aligned diagnosis.

The real cost of scaling AI

The debate opened on an uncomfortable but necessary note. Investing in AI brings costs that companies don’t always see coming — more infrastructure, more data control and an organisational transformation that goes far beyond technology. As the PwC partner put it, the level of maturity is “still more than embryonic”, and the challenge is to truly quantify the return on investment. Multiverse added the other side of the coin. “The key word is control”, but control that comes with independence and sovereignty. Poorly managed, AI becomes a cost before it becomes a benefit.

Our view: generic AI doesn’t know your company

At Logixs, we made the case for where, in our experience, the problem really starts. Generic AI knows a lot about the world and nothing about your company — it doesn’t know your products, your rules or who makes the call in each process. That’s why the pilot shines in the demo and stalls as soon as it meets the reality of the business.

Francisco explained the alternative on stage. “We need to find all the knowledge that exists within the company and capitalise on it in a contextual engine, a brain you can connect to any model that knows what your company does, why it does it, your policies and your rules. That’s when we’ll move to the next level of artificial intelligence.” It’s exactly the principle we built ADA on.

And before any of that, there’s one condition we repeat at every forum — governance. “Anything that isn’t well defined and well controlled won’t make it to production”, Francisco warned. Far from slowing things down, clear boundaries are what accelerate adoption — only when teams know how far they can go does real internal innovation begin.

Overcoming fear (through training)

There was also consensus on the human side. “The key to everything is overcoming fear, and we beat it with training and culture”, Francisco argued. On the impact on jobs, his favourite metaphor made a return. “AI is a rocket, but to get to Mars someone has to fly it” — junior profiles will need to be able to create and build, not just execute. And on the organisational front, full agreement with PwC that adoption has to be led from the top. “It’s essential that it’s led by senior management”, Francisco concluded.

Three takeaways

The panel closed with a summary we fully agree with — cost control, sovereignty and overcoming fear. Investing in AI is no longer enough. It has to be governed, measured and built on protected data to become a reliable, auditable and profitable tool.

Thank you to El Confidencial and PwC for organising, and to Multiverse Computing for the conversation. You can read the full report on the debate here . And if your company is at that point, with pilots that never reach production and costs that don’t add up, let’s talk.