How is your data stack changing in 2025

Hey All,

New Year, new stack, maybe… …how are you all changing around your data stacks this year? :thinking:

  • Leaning into GCP’s LLMs – if so, please tell me your experience in the comments
  • Deploying a renewed (or new) commitment to governance? Is so are you doing it manually or with tools?
  • Do you use other tools with your Looker like dbt for modeling or Rollstack for automating recurring reports?
  • Are you going to Google Cloud Next? If you’ve gone in the past what did you get out of it?

Thank you all for making the community such a helpful resource for analytics and BI!

3 Likes

I think governance wins the day this year. We all know garbage data and compliance issues hold back good AI and better reporting.

3 Likes

All things AI, but yes governance and data quality is the name of the game. Also automate all the reports! :joy:

2 Likes

Yes for Google NEXT, first timer.

We witnessed a lot of large enterprises aggressively adopt AI in 2024 – 2025 is a year where commercial size companies, and small enterprise will do their utmost to catch up.

In terms of the data stack it’s all about trusted AI, governance, cost-effective storage, and continuing to automate, rollstack for automating reports, dbt still very good for modeling.

Educating more people about the data stack continues to be important.

1 Like

So it sounds like governance, AI, and automation with tools like Rollstack and dbt are winning the day this year?

Are there any resources you can recommend where I can learn more about putting systems of data governance in place?

1 Like

Yes, there’s literally an O’Reilly book on data governance – I’d start there. One hack I do with O’Reilly, is follow and connect with the authors on social media (LinkedIn) too.

1 Like

Oh nice. I haven’t read that one yet. “Designing Data-Intensive Applications” is a classic. I haven’t tried dm’ing O’Reilly authors yet. That should be a hoot.

Well I posted this before Gemini 2.0, ChatGPT 4.5 and deepresearch, Grok3, and DeekSeek – any updates on the AI front from anyone?

I wouldn’t trust my data with deepseek. Claude’s engineering capabilities are so so good.

1 Like

I’ve found that Gemini’s Flash 2.0 is very very good, and for the less technical team the fact it can do some analysis in Google Sheets is a plus.

Have you all come up with solid prompts that get better answers and reduce hallucinations?

Now that we’re deep in Q2 2025, and we’re past Google Next, anything new?