Data Decoded 2025: Key Takeaways and Maia Highlights

We’ve just returned from Data Decoded 2025 at Manchester Central – and it was a proud moment for Matillion to showcase Maia in our home town. 

In its first year, the event brought together the UK’s brightest data professionals, engineers, and business leaders to decode the future of AI, analytics, and data engineering.

From a buzzing stand packed with demos, to standing-room-only theatre sessions, Data Decoded gave us the perfect platform to show the power of the world’s first agentic data team in action.

Let’s recap some of the highlights from the event…

Maia in Action: AI Needs Data, Data Needs AI

Our stand at Manchester Central quickly became one of the busiest corners of the show floor. Attendees gathered throughout both days to see Maia – our agentic data team – tackling complex data engineering tasks live.

We ran back-to-back demos showing how Maia can:

• Author complex pipelines using natural business language
• Build and test connectors in seconds
• Debug and optimize pipelines in real time

The energy at the booth was unmistakable, with visitors eager to understand how agentic AI can unlock unparalleled data productivity gains.

As one attendee put it: "Some of the tasks that take people a couple of months could be done in a few minutes. With Maia, I would be able to do a lot more in the day. The sky's the limit."

Standing Room Only: Matillion Theatre Sessions

Matillion hosted two packed-out theater sessions during the event, each exploring how agentic AI is redefining what’s possible for modern data teams.

Leveraging Context Files to Build Data Pipelines in Line with Governance Best Practices

In this session, Joe Herbert, Maia’s Principal Solution Architect, demonstrated how Maia eliminates the bottlenecks slowing BI, Data Science, and Analytics teams. Acting as an agentic data team, Maia translates natural language into clean, production-ready code while ensuring governance and auditability. Attendees left with a clear picture of how agentic AI can multiply productivity and keep pipelines compliant with organizational standards.

Key takeaways included:

• How agentic AI removes data engineering bottlenecks
• Generating governed pipelines from natural language with Maia
• Achieving up to 100× productivity gains

Future of Data Engineering in an Agentic World

Matillion’s Mike Harms, Manager, Field Engineering, and James Peckover, Senior Technical Account Manager showcased a live Maia demo, alongside roadmap teasers. Attendees saw how Maia converts natural language prompts into YAML-based, human-readable Data Pipeline Language (DPL) to generate graphical pipelines, interact with Snowflake metadata, troubleshoot in real time, and even build custom connectors in seconds.

Key takeaways included:

• How agentic AI can accelerate data engineering workflows
• Transforming natural language prompts into YAML-based pipelines
• Real-time debugging and metadata interaction capabilities

Both sessions drew enthusiastic crowds, sparking conversations that carried on well beyond the theater. 

Executive Roundtable: What Data Leaders Need to Know Now

Matillion’s CEO Matthew Scullion hosted the event’s most insightful gathering of data leaders. The exec roundtable 100x Data Productivity: How Agentic AI Transforms Data Strategy and Eliminates the Engineering Bottleneck invited data leaders from across industries to discuss the agentic future.  

The lively conversation unpacked how AI is modernizing ETL, consolidating fragmented tech stacks, and delivering unprecedented productivity gains. The session sparked candid discussion on how organizations can operationalize AI securely and at scale.

Keynote: The Era of Agentic AI

Day 1 closed with an urgent and important keynote from Matthew Scullion, setting the tone for the entire event. In “The Era of Agentic AI”, Matthew explored how data leaders and their teams need to embrace AI now, or risk falling behind. 

“AI isn’t hype – it’s real, it’s here, and it’s already rewriting how we work,” he told a packed audience. “But to make it real, we need data teams who can operate at machine scale.”

Matthew spoke about the imbalance at the heart of today’s enterprise: a world with machine-scale demand for data but human-scale capacity to deliver it. The only way forward, he argued, is for AI to come back and rescue data – transforming data engineering from a bottleneck into a force multiplier.

He explored what that means for teams today:

• For data leaders – enabling AI to safely modernize ETL and eliminate bottlenecks.
• For engineers – using agentic AI to amplify their output 30–50x.
• For businesses – unlocking self-service data capabilities that were previously impossible.

The keynote ended with a live demo of Maia on stage – showing the world’s first agentic data team tackling a complete pipeline from a natural language prompt, in minutes.

“AI needs data. Data needs AI. A machine-scale problem demands a machine-scale solution,” Matthew concluded. “That’s what Maia is – a new way of working that turns backlog into breakthrough.”

The message was clear: the future of data productivity is already here – and it started in Manchester.

Looking Ahead: The Future of Data Productivity

If Data Decoded proved anything, it’s that the data industry is entering a new phase – one powered by agentic AI. The conversations in Manchester reflected a growing consensus: the next competitive edge won’t come from more people or more tools, but from more intelligence built into every layer of the data stack.

Maia represents that shift. It brings engineering precision, business context, and AI-driven speed together – enabling teams to operationalize AI safely and scale with confidence.

Ready to take the keyboard and see Maia at work in your own environment? The next step is a Maia Session: 60 minutes to pit Maia against your toughest data challenge. 

Missed Big Data London this year? Catch up on the highlights from the most talked-about theatre sessions in our webinar From Legacy to AI: Redefining data engineering for the agentic world.