Why Goldman Sach's latest AI discussion hits the mark!
- Chelsea Wilkinson
- Dec 2, 2025
- 2 min read
Every now and then, something cuts through the AI noise and gets right to the heart of what actually drives value. This Goldman Sachs discussion does exactly that.
What struck me most is how clearly it reinforces something we see every day at DataDiligence: AI isn’t magic. It’s 'momentum' built on top of clean, connected, well-understood data.
Without that foundation, even the most powerful models can only go so far. And yet, in so many mid-market organisations - especially those backed by private equity - the most valuable data lives in pockets, Excel spreadsheets, legacy systems, and “we’ll sort it later” workflows.
That’s the real opportunity.
The conversation, led by Goldman Sachs’ Neema Raphael, underscores that the next wave of competitive advantage won’t come from access to models (everyone will have them) but from unlocking the proprietary data sitting inside your business, giving teams the ability to make faster, sharper decisions with confidence.
It also highlights a point I’ve been making recently: AI becomes genuinely transformative only when paired with strong governance and processes, shared definitions, and a single source of truth. Without that, you’re building on sand.
So why am I sharing this?
Because it’s refreshing to hear a major institution articulate, in plain language, what DataDiligence has been championing for years — that data quality, structure and business context matter far more than the hype cycle. It’s always been the unglamorous part of the story, but it's absolutely the lever that separates incremental improvement from real enterprise value.
And this really matters on the investor side, too.
In our Data & AI Due Diligence work, we see how often the investment case rests on assumptions about data that haven’t been tested — claims about customers, transactions, AI potential, automation, or “platforms” that sound great until you actually look under the hood. Proprietary data, processes, and structure can strengthen an investment case… or dilute it just as quickly. Which is why testing what’s real - and what’s missing - is now essential.
If you’re thinking about where to start with AI, or how to make your existing (or potential) investments actually deliver, this is well worth a read. It’s a timely reminder that the organisations who treat data as a strategic asset - not a by-product - are the ones who will pull ahead.
Comments