# Integrating AI for data analysis

**URL:** <https://forum.inductiveautomation.com/t/integrating-ai-for-data-analysis/113376>\
**Category:** Ignition\
**Tags:** perspective\
**Created:** [January 17, 2026, 5:23pm UTC](https://forum.inductiveautomation.com/t/integrating-ai-for-data-analysis/113376 "2026-01-17T17:23:22Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![Aiasa21](https://avatars.discourse-cdn.com/v4/letter/a/aeb1de/32.png) [@Aiasa21](https://forum.inductiveautomation.com/u/Aiasa21)\
**Post date:** [January 17, 2026, 5:23pm UTC](https://forum.inductiveautomation.com/t/integrating-ai-for-data-analysis/113376/1 "2026-01-17T17:23:22Z")

</div>

Good evening,

to do some testing, I implemented a call from a view to Gemini.

Specifically, I added a table in a view with default values (population data, cities, etc.), and through a textbox I ask Gemini for information, passing the table as the data source. Gemini’s response is then displayed in a markdown component.

So far everything works correctly, but it’s clear that this opens up a lot of possibilities in terms of what can be analyzed and how the response can be obtained, for example in tabular format or as a chart.

I was wondering if anyone else has already tested something similar and, if so, whether they have any advice on how to approach this in the best way.

Thanks.

---

<div class="post-metadata">

**Author:** ![pablo.maestre](https://avatars.discourse-cdn.com/v4/letter/p/49beb7/32.png) [@pablo.maestre](https://forum.inductiveautomation.com/u/pablo.maestre)\
**Post date:** [January 27, 2026, 7:27am UTC](https://forum.inductiveautomation.com/t/integrating-ai-for-data-analysis/113376/2 "2026-01-27T07:27:01Z")

</div>

Good morning Aiasa21,

I'm currently battling with IT to get approval to use Gemini. It's the usual problem: organizational data. They are digging their heels in on that, while we miss a huge opportunity to analyze data like never before.

I have run a lot of tests; I think the most relevant ones involved passing results from **Named Queries** —pulling information from our MES and GMAO (Maintenance data) for a full week— to identify improvement opportunities. It still has some room for growth, but it saves us all the work of documentation, selection, and resource allocation when investigating our top 3 breakdowns

I also use Markdown for the responses. Current LLMs have the capability to build it perfectly on the first try, and it looks visually great.

Thanks. Pablo.