TL;DR: Everyone has/is/does “AI.” Governance is the biggest topic. BI 2.0 is no longer interesting.

Two weeks ago, I spent a day-and-a-half at the BigDataLDN (BDL) conference in London to get the temperature of the technology and sentiment of the data world. BDL is one of the biggest conferences for data in the UK, with 154 exhibitors, 15,000 attendees, 15 theatres and 300 speakers. It’s an important part of the data conference calendar.

What Did I Find?

Before digging into more specific exhibitor and speaker insights, here are the big points worth mentioning from this conference:

• It’s a small world. Many of the same faces and characters are still part of this solar system. I spoke to people like Francois Ajenstat at Golden Analytics, Steve Morse of Neo4j and Steve Prok, who is ex-Tableau, along with current clients and former InterWorkers. Speaking of clients …
• Meeting our clients is invaluable. Conferences are always great for this. I was fortunate to sit down with some InterWorks clients to hear what their current challenges are, as well as which tools they’re looking to for the future. Topics like governance, aligning global teams and which tools to use for specific functions like ETL were all discussed.

Exhibitor Messaging

The following insights come from my time perusing the Exhibitor Hall. It’s always interesting to contrast messaging from exhibitors – most often compromised of vendors and consultancies with something to sell – with that of conference speakers, who are generally the organisations they’re targeting. Here’s what I saw from exhibitors:

• Can we trust what we are seeing? The core message is the same everywhere. The difference is the technology deployed. Semantics and the context layer seem to be where the big bunfight is, and there are so many tech vendors that look very similar. People like Atlan are rebranding as agent infrastructure with governance ties. This also extends into quality and observability, and 40 exhibitors talk about it.
• “AI” feels like the “Big Data” of this era. Everyone is talking about it. 95 of the 154 exhibitors have the word “AI” in their title or headline, and 55 of them use the terms “Agentic AI” or “Agentic Analytics.” As we know from past movements, the ultimate validity and authenticity of these statements may vary from vendor to vendor. AI is undoubtedly a meaningful movement right now, but it’s also subject to the same marketing hype we once saw with things like “Big Data” 10-15 years ago.
• Hyperscalers are filling in. Google, Microsoft and Oracle are adding data integration and replication to their suites.
• Claude and MCP. Claude is mentioned often, and so is MCP, which connects your initial access surface to your other tools.
• BI is repositioning as “agentic analytics.” ThoughtSpot, Pyramid, Domo and Omni all lead with AI adoption with self-service. Sigma is the most differentiated play of this group, with its “runtime layer for analytics, apps, and agents on live data.” Dashboards are very much a thing to move beyond.
• Cost control is a smaller topic. 18 exhibitors talked about warehouse spend, including Zipher, Altimate, Greybeam, Rabbit, Flexera and MotherDuck. I assume they annoy the Snowflakes and Databricks of the world.
• Platform consolidation is noticeable. Fivetran and dbt Labs are the standouts here. To be honest, I can see many of the above companies either disappearing or consolidating by next year.
• There were more this year, but all with exactly the same message: “we make AI real.” That included CI&T, CACI, Flatiron, 7DExperts, Biligence, Advancing Analytics, Celebal and Datapao, a Databricks SI that worked for a long time as a Databricks subcontractor.
• What was absent? Tableau and Looker had zero attendance and zero mentions. Qlik, Salesforce and Alteryx were mentioned as connectors rather than core tech. Out of the tech vendors, I honestly think Sigma Computing had the most interesting message and differentiator against a sea of sameness.

Speaker Messaging

I didn’t attend every session, so this section comes from the BDL speaker synopses, summarized with help from Claude:

• Context is the new semantic layer. There was a new “AI Context” theatre, and around 25 sessions covered context layers, ontologies, knowledge graphs and semantic models as what agents rely on. Examples include Atlan’s “WTF is the Context Layer” keynote, Coginiti’s “Context Stack,” Neo4j’s “Knowledge Layer,” PostHog’s “Context Warehouse” and Fivetran’s “Open Context Layer.” An AWS pharma case study, “Your Semantic Layer Is Your AI Strategy,” went from text-to-SQL hallucinations to 2,500 users.
• Pilot to production, and the value gap. A large group of talks asked why AI stalls after the pilot. Examples include Softwire’s “12% Club,” Prospore’s “Value Problem,” Dataiku’s “Dependency Trap,” Cynozure on scaling agentic AI and Coca-Cola Europacific Partners with Deloitte.
• Post-dashboard BI. This was the sharpest shift for our world. Compare the Market “ditched the dashboard,” Omni argued “the best dashboard might be no dashboard,” Betsson ran a “post-dashboard stack,” Entain said “a chatbot on your dashboard won’t save you” and ServiceNow called it “the end of read-only analytics.” Hex, Omni, Lightdash, ThoughtSpot, JetBrains and Pyramid (now part of ServiceNow) all pushed agentic or personal-agent analytics.
• Governing agents. Several sessions covered observability, agent sprawl and accountability: Bigeye’s “When Agents Outnumber Humans,” Snowflake’s “Who Gets the Blame When Agents Talk to Agents,” Dataiku on governing actions that can’t be undone, Redpanda and Acceldata. The EU AI Act “control cards” talk also sits here.
• Talks came from HCLSoftware/Actian, Dremio, Cloudera, Telefonica Tech, BAE Systems and SERPRO. Speakers framed sovereignty as who stays in control, rather than where data is stored.
• Token economics. Snowflake’s opening “Valuemaxxing in a Tokenmaxxed Game,” ThoughtSpot’s “Tokenomics without panicking the CFO” and Databricks on “Context, Control, Cost, and Choice.”
• Architecture convergence. OLTP and OLAP are merging (Databricks, ClickHouse). Streambased says “AI is killing ETL.” Starburst talked about querying data where it sits instead of migrating it, MotherDuck about agents as heavy users of the data stack and others about the death of batch processing.
• Agents doing the data engineering. Matillion, Coalesce, QuestDB and Factory all covered agents building and fixing pipelines, and the data engineer becoming an agent manager.
• People and culture. Two theatres covered data culture, talent gaps, the changing analyst role, data leaders as transformation leaders and Data for Good. Louis Theroux closed day two on “Truth in the Age of AI.”

Customer Stories Worth Noting

Dunelm, Liverpool FC, Shell, JLR, Experian, Lloyds (three sessions), Monzo, Expedia, BBC, Netflix, Wise (moving off legacy BI), ASOS, Kingfisher, Royal London Asset Management, Sagacity, Mondelēz and Haleon were all incredibly interesting stories. It might be worth perusing LinkedIn or looking up to see if those stories are present somewhere digitally as they’re doing some interesting work.

How Does This Impact InterWorks and Our Clients?

Our goal at InterWorks has always been to build trust in the data used by people making their business decisions. It appears the market has realised this is the most important change we can be managing in organisations. It’s good to see, but the problem now is which technology should I choose to help me with this? My answer, use what you’ve got and do the hard work of building out the knowledge around your data assets. Create the metadata, create the semantic layers and create the context that will build your trust in the answer you get.