For the second year in a row, I had the privilege of attending the Databricks Data + AI Summit in San Francisco.

For the second year in a row, I had the opportunity to stand on stage and speak to a packed room about something I’m incredibly passionate about, helping organisations unlock value from their data.

Looking around the room, one thing became very clear. The excitement around AI hasn’t slowed down compared with previous year. If anything, it’s accelerating.

Every keynote was bigger, every announcement more ambitious, and every conversation centred around the same topic: AI is moving from experimentation to enterprise reality. But after four days of product launches, customer stories and technical deep dives, I found myself thinking about something much simpler.

Behind every announcement is a business problem waiting to be solved. And that’s the part I believe matters most!

We don’t need more AI.

We need better business outcomes. It’s easy to become distracted by new products.

  • Agent Bricks.
  • Genie One.
  • CustomerLake.
  • LTAP.
  • Lakehouse RT
  • Unity AI Gateway.

On paper, they’re all impressive.

But technology has never been the goal. Business value is.

The organisations that succeed over the next few years won’t necessarily be those using the newest AI models. They’ll be the ones solving the right business problems.

  • Can your marketing team personalise every customer interaction without doubling headcount?
  • Can your finance team answer strategic questions in minutes instead of waiting days for reports?
  • Can your operations teams identify production issues before customers even notice them?
  • Can your engineers spend more time building products instead of maintaining infrastructure?

Those are the conversations executives should be having.

AI is becoming operational

One message appeared repeatedly throughout the Summit.

We’re moving beyond AI pilots. AI is becoming part of everyday business operations.

Take Genie ZeroOps.

At first glance, it looks like another operational feature. In reality, it addresses one of the biggest scaling challenges organisations face. Every new AI model, pipeline or application increases operational complexity. Eventually, engineering teams spend more time maintaining platforms than improving them. AI monitoring production systems, investigating incidents and proposing fixes isn’t simply an engineering productivity feature. It’s about reducing operational risk and increasing business resilience.

Every minute a production pipeline is unavailable can delay customer communications, disrupt supply chains or affect executive decision-making.

Reducing downtime isn’t an IT metric. It’s a business metric.

AI should empower the business, not create more work

One of my favourite announcements was Genie One.

Not because of the technology. Because of who it’s built for.

For years we’ve invested in self-service analytics. Yet many business users still depend on data teams to answer relatively simple questions.

  • “Why did sales decline in the North?”
  • “Which products are driving margin?”
  • “How many customers haven’t purchased in six months?”

Each request enters a backlog. Someone writes SQL. Someone builds a dashboard. The business waits.

Genie One changes that relationship: Instead of asking the data team for information, business users interact directly with governed enterprise knowledge. The outcome isn’t fewer data engineers. It’s more strategic data engineers. Less time producing reports. More time solving business problems.

Personalisation finally becomes economically possible

Another announcement that really stood out was CustomerLake.

Marketing has always wanted true one-to-one personalisation. The challenge was never ambition. It was cost. No organisation can assign a dedicated marketer to every customer.

So businesses compromise. Segments. Campaigns. Broad assumptions.

CustomerLake fundamentally changes that equation.

Instead of one campaign for thousands of customers, AI agents continuously determine the right message, through the right channel, at the right time for every individual customer.

  • For retailers, that means higher conversion.
  • For banks, more relevant financial products.
  • For insurers, proactive engagement before renewal.
  • For customers, experiences that finally feel genuinely personal.

Governance is no longer optional

Perhaps the strongest message from the Summit wasn’t about AI at all.
It was about governance. Unity Catalog. A few years ago no one was excited about it.
As organisations move from a handful of AI assistants to potentially thousands of AI agents, governance becomes the foundation that allows innovation to scale safely.

Governance isn’t there to slow teams down. It’s what gives organisations the confidence to innovate faster. Because once security, permissions, lineage and policies are inherited automatically, teams spend less time navigating compliance and more time delivering value.

That’s a significant mindset shift.

Governance moves from being a control mechanism to becoming an accelerator.

Before AI comes trusted data

One thing hasn’t changed. Every AI capability announced at Summit depends on one thing.
Trusted data.

Whether it’s Genie One understanding your business…CustomerLake personalising customer journeys or Agent Bricks building enterprise AI agents…
they all assume the same thing:

  • Your organisation knows what a customer is.
  • What revenue means.
  • Which data is trusted.
  • Who is allowed to access it.

Without that foundation, AI simply produces inconsistent answers faster. The future of AI isn’t about bigger models. It’s about better data. And of course, Genie Ontologie is making all this possible.

My biggest takeaway

If I had to summarise this year’s Summit in one sentence, it would be this:

Competitive advantage will no longer come from experimenting with AI. It will come from operationalising AI responsibly.

  • That means building on trusted data.
  • Embedding governance from day one.
  • Starting with business outcomes rather than technology.
  • And enabling people across the organisation, not just engineers, to make better decisions.

Because ultimately, AI isn’t the destination. Better business outcomes are.

I left San Francisco more convinced than ever that we’re entering a new phase of enterprise AI. Not one defined by technology alone, but by organisations that can successfully combine trusted data, intelligent platforms and clear business purpose.

And that is an incredibly exciting place to be.

See you next year!