Data 2030 Summit
Last week I attended the Data 2030 Summit in Stockholm, a data strategy conference and round-table event organized by Hyperight. The event was divided into three tracks: Data Strategy and Governance, Modern Data Platform and Data Architecture and Quality.
The event gathered 262 attendees, from 121 companies and 22 countries, and covered a broad set of industries, including financial services, manufacturing, transport and logistics and healthcare.
My Personal Takeaways:
Be Obsessed with Value: GenAI has passed the peak of inflated expectations in the hype cycle, and there is now a big, and sane, push towards finding use cases that deliver business value instead of focusing on (Gen)AI as the new shiny thing. Making your data available, while keeping information secure and of good quality is necessary for almost every use case. Prepare for both human and AI consumers from the get go.
Democratize your Data: Most companies are on a path towards democratizing the data, very few are completely there. To find the balance between centralization and siloes, most seem to be aiming for a hybrid operating model, in some cases according to the data mesh principles, in other cases some variation of it.
Focus on People and Culture: As always, when changing technology you can’t forget about the people and culture. Bridging the gap between business and tech is key to succeed in this area, make sure you invest in the people who are good at speaking both languages, translating business needs into tech requirements and vice versa.
Some Architectural / Tech Trends:
Agentic AI - Thinking Fast and Slow: If GenAI has moved over the peak in hype, Agentic AI is now on the rise. The analogy goes to Kahneman's Thinking, Fast and Slow, where agentic AI promises to combine both system 1 and system 2 thinking.
Graph Database + Master Data: Using a graph database and an ontology to work with master data, as described by Handelsbanken, seems like an interesting approach.
Vector + Retrieval-Augmented Generation (RAG): Vector databases, combined with RAG is becoming a common pattern, where you combine the benefits from using a foundation model/LLM together with your own data.
Operational and Analytics Data Converging: With a modern, often cloud based, data platform analytics use cases and operational ones are converging. Data products published on the platform don't necessarily need to be used only for reporting and analytics, they can also be used by other systems, for example in ML and AI implementations.
One Size Doesn’t Fit All: It was nice to see examples of very different strategies and implementations, everything from completely open-source based platforms, to using a single vendor’s Saas platform. Depending on the requirements and context both ways work just fine. One size doesn’t fit all, after all!
All in all the conference was a nice experience. I especially liked that many speakers were sharing both successes and failures, from the real world. The mix of round table discussions and more traditional key notes and talks was good. I attended the conference with a group from my current customer, so it was also a really good opportunity to discuss things from our different perspectives, coming from different parts of the org.
The strong focus on realizing business value through data and AI, and on bridging the gap between business and tech, is very much aligned with what we are trying to achieve with Trice.
