We went on the World Tour and this is what we learned ??

With the AI frenzy turning up to 11 on the Databricks World Tour, Kris and Mithali attended the Melbourne conference to (find out how quickly we'll all be replaced*) get the latest in data, analytics and AI on the Databricks Lakehouse.

Here are three key takeaways from a day of wall-to-wall AI.

1 - Generative AI is an emerging opportunity to bring data to all users in your organisation

• Databricks is all in on opening its platform to more users. Its assistant can already generate code through natural language as a great productivity tool.
• With the announcement of LakehouseIQ, the gates will widen, allowing end users lacking a technical background to explore data, without the need for SQL or Python – just using natural language.

2 - But with the buzz of Generative AI... you can’t just stick a LLM on your data and expect it to work.

• Much focus of the summit, and rightly so, was on getting the basics right. Bringing your data together in one place, making sure it is organised, cleaned and secured correctly.
• It was great to see one of our customers, Alinta Energy, talk about this exact need, and the fantastic work they’re doing bringing all their capability into the Databricks Lakehouse to create a single, organised, accessible foundation.
• Other customers discussed the need for MLOps, having solid operational processes to manage your ML and AI capabilities – all within the same solid foundation with your data.
• If you don’t have your data and systems integrated, organised, accessible, and importantly, of good quality, your LLM - let alone any other ML model - will miss the mark and not provide the value you were looking for.

3 - Generative AI is here to amplify and accelerate, not replace.*

• A DEEP sigh of relief!
• Responsibility around AI was another area touched on by Steve Nouri during the talks. There are some areas where these tools will just not be appropriate or need to be used with caution – in most cases, the output needs to be reviewed and used in context… be careful just presenting the content without checking its references!
• Overall we are a long way from removing the need for skillsets in data and analytics. Current and future applications of generative AI will exist to improve productivity – and make data more accessible to the masses. Maybe one day, but for now, it relies entirely on the user having a solid foundation of what to (precisely) ask and how to apply it.

Don't be a stranger! Reach out to Kris or Mithali for any questions.