LinkedIn Post: The Scattered Data Problem & AI Readiness¶

Superseded 2026-08-18. This post is now AI for Wealth Managers #4, scheduled 19 Aug 2026. The live version, with the date fix and the editor notes, is at content/posts/04-your-ai-is-only-as-good-as-your-data.md. Kept here for provenance — don't post from this file.

Date finalised: 2026-06-12 Status: Ready to post Type: Problem-agitation / genuine inquiry post Positioning angle: Scattered data makes downstream AI useless — fix the foundation first. No meeting prep reference (deliberate — this post is purely about data being hard to find and AI readiness).

Final Copy¶

Wealth manager, are you being bombarded with new "life-altering, business-rescuing" AI solutions right now? Note takers, report writers, agentic solutions and "copilots" for everything. The list is endless.

Here's the uncomfortable part: almost none of it will work as promised. Not because the AI is bad, but because of where your data lives.

A few weeks ago, there was a media frenzy about Anthropic's new Claude skills for financial services. Apparently, they're wonderful. World-changing, even.

My honest take: those skills are worthless sitting on top of poor-quality data. And by quality, I mean accurate, available, and up to date.

Because here's where a wealth manager's data actually lives:

• In a CRM (if you have one). • In a shared folder structure that has evolved over many years. • In old email conversations. • In handwritten notes from the last 5 meetings. • In WhatsApp and other disconnected comms channels. • And the most important parts are in someone's head.

Portfolio data? Locked in PDFs, or spread across platforms where every one has a different report structure and a different API — and some barely give you a login screen.

AI is only as good as what you feed it. Point a reporting AI at data scattered across six places and three formats, and you get confident-sounding reports built on a fraction of the picture. You might as well not add the AI at all.

Which brings me to the thing I can't get past: I've been building fintech for wealth managers for more than 25 years, and this scattered data problem is older than my career. Everyone in the industry knows about it. Smart people have tried to fix it. It's still here.

My best theories:

• Integrating the systems is brutal. Every platform and CRM is different, and the data never quite matches. • Capturing it is high-friction. Meeting notes, CRM updates, compliance documents — they all compete with actual client work, so they get done late, in a rush, or not at all. • And the many systems don't even agree with each other. Move money between two of a client's accounts and it can vanish from one before it lands in the other.

So, my question is: "How do we properly prepare a wealth management business to take advantage of AI?"

I obviously have strong views — I'm building in this space. But I genuinely want to know:

If you've solved the data problem end to end, I want to hear how you did it. And if you tried and gave up, what killed it?

Strategy Notes¶

Structure: Call-out hook → contrarian take on AI hype (Claude financial skills media frenzy) → data inventory list (recognition engine) → garbage-in-garbage-out argument → 25-year credibility + "why unsolved" → three theories → big question → comment prompts.

Why no product mention: Post works as genuine inquiry. The "I'm building in this space" disclosure is the only reference — disarms bait cynicism without becoming a pitch. Comment section doubles as market research (war stories = ICP self-identifying).

Comment-section ammunition (deploy in replies, not the post): - LLMs finally make unstructured data (notes, emails, WhatsApp) usable — the reason this is solvable now when it wasn't before - Integrations were never economically viable for a single vendor to build speculatively - Platform incentives favour lock-in, not interoperability - "Don't buy AI tools — get AI-ready" reframe

Do not: pitch ClientIQ in early comments; answer the big question yourself in the post; mention meeting prep or transcription.

Fact basis: Anthropic Claude financial services skills media coverage (Feb 2026 Claude Cowork plug-ins triggered enterprise software selloff per Fortune; May 2026 financial services skills additions). Stock angle deliberately removed from final copy — hype framing only.