# Is Your Data Training Your Future Competitor?

> Australian healthcare and aged care providers are pushing operational knowledge into AI tools they don't control. That knowledge is a strategic asset.

Somewhere in your organisation this week, someone pasted an incident report into a chatbot.

Picture the person. A quality manager at an aged care provider, three notifications behind, drafting a serious incident report before the 24-hour clock runs out. She pastes the incident notes into a public AI tool and asks it to tidy the wording. It helps, so she pastes the next one. Within a month the tool has seen how her organisation classifies incidents, which phrasings get through the regulator without follow-up questions, and how staff describe what happens on the floor at 2am.

None of that is a data breach. The names were probably stripped. But something still left the building: the judgement.

For healthcare and aged care providers, data is more than a collection of records. It is operational memory. It captures how care is delivered under pressure and how the organisation stays viable under regulation. That knowledge took years to build, and it is what separates a provider that works from one that struggles.

Feed it into AI systems you don't control, without a deliberate strategy, and you are doing more than using a tool. You may be training the layer above you.

## What happened to Figma

This risk is not specific to healthcare. The clearest recent example comes from software.

In 2026, Anthropic launched Claude Design, a product that moved directly into the design workflow space where Figma operates. Anthropic's chief product officer had served on Figma's board until shortly before the launch. Figma's market cap, listed at US$11.27B by CompaniesMarketCap in July 2026, sits well below its post-IPO highs.

To be fair to everyone involved: none of this proves misuse of Figma's data, and a market re-rating has many causes.

What it does show is that model providers are no longer staying in the infrastructure layer. They are not content to supply models for others to build on. They are building the applications and workflows themselves, and they are better placed to choose where to build when they can see repeated workflow patterns. A model provider with enough visibility into an industry's workflows and repeated decision points can work out where the value sits. It doesn't need your database for that. The shape of your work is enough.

![Figma market capitalisation chart showing a steep fall from its post-IPO highs to US$11.27B in July 2026](/images/figma-market-cap-dive.png)

*Source: CompaniesMarketCap, July 2026.*

Investors appear to be pricing in that risk quickly. Once the market starts asking whether the model provider underneath a platform is a supplier or a future competitor, the platform gets re-priced. The same question applies to a care provider feeding its operational patterns into tools owned by companies that are moving up the stack.

Back to the quality manager. Nothing she pasted was personal information, and no privacy review would flag it. What she was sharing, one report at a time, was the way her organisation thinks. Why that kind of leakage slips past every privacy framework you already have is the next post in this series.

I work with regulated organisations on exactly this problem. There's more on the [about page](/about/) if that's useful.

---

Source: https://kads.au/writing/is-your-data-training-your-future-competitor/
Author: Kads Aziz