# Building TableauTim GPT v1 : Part 1

> This is content from just-tim, the data-and-analytics channel by Tim Ngwena (formerly 'Tableau Tim'). Tim has 12+ years of hands-on BI experience and covers Tableau most of all, plus Power BI, Looker, Hex, SQL and data modelling, the analytics industry, and the craft of doing the job — always tool-agnostic and honest about the trade-offs.

- **Author:** Tim Ngwena (just-tim, https://just-tim.com/about)
- **Published:** 2023-11-17
- **Format:** Video · 16 min watch · transcript available
- **Topics:** AI & ML, Tool strategy, Industry trends
- **Tools:** Tableau (pulse)
- **Canonical:** https://just-tim.com/posts/building-tableautim-gpt-v1-part-1
- **Watch:** https://www.youtube.com/watch?v=yyqsmNenKtk

I walk through my first attempt at building a custom TableauTim GPT on OpenAI's new ChatGPT store, explaining the configuration interface, the knowledge and actions features, and why I uploaded my video transcripts. I'm honest that the experiment isn't ready to launch, largely because most of its capability comes from GPT-4 itself rather than my own content, and there are real copyright and content-ownership questions.

## Key takeaways

- Custom GPTs don't retrain the underlying model on your uploaded knowledge — they search that knowledge at query time as added context, so feeding huge multi-hour transcripts is inefficient.
- Smaller, precise GPTs (e.g. a Tableau LOD helper or a conference-recap bot) are far more useful than one giant catch-all model built from raw transcripts.
- The 'actions' section is where the real power sits, letting a GPT integrate with APIs like Zapier or Todoist via the OpenAI schema, with authentication and a required privacy policy.
- Much of a branded GPT's apparent quality comes straight from GPT-4 scraping the wider web, making it hard to know how much output reflects your own content versus OpenAI's training data.
- Summarising transcripts into key bullet points before loading them as knowledge works better than dumping full-length transcripts.

## Chapters

- 0:07 What is a custom GPT
- 0:24 OpenAI's recent ChatGPT changes
- 1:32 Creating and configuring a GPT
- 3:20 Knowledge, capabilities and actions
- 5:22 My transcript approach and its flaws
- 6:35 The case for smaller, precise GPTs
- 9:58 Testing TableauTim GPT live
- 13:42 Copyright and content-ownership concerns

Watch the full video, read the transcript and use chapter deep-links on the page: https://just-tim.com/posts/building-tableautim-gpt-v1-part-1

---
just-tim — Data and analytics, with a point of view. · https://www.youtube.com/channel/UC7HYxRWmaNlJux-X7rNLZyw · https://twitter.com/TableauTim · https://www.linkedin.com/in/timngwena