I pride myself on being CRM-agnostic. I believe that any CRM, when properly configured, can be used by an advisor to serve their clients well. But I am getting a little nervous for the firms who are building their entire practice on a foundation of artificial intelligence. Here's why...
All CRMs are databases that are designed to hold prospect and client details and make them available for query — whether by old-school deterministic SQL statements, or via probabilistic LLM prompts. As long as your data is clean and your query doesn't suffer from drift or hallucination, the results are going to be roughly the same.
However, I am seeing a change in attitude from both the general public and politicians when it comes to AI. There is now considerable pushback from communities that don't want data centers built in their backyard and political will that is beginning to bend towards regulating firms and holding models accountable for misaligned output. And that writing on the wall is what has me concerned.
Three pressures that could raise the cost of AI
- When the public doesn't want you to build, your cost to build goes up due to the increased incentives required to placate the public.
- When regulations begin to roll out — and I believe we'll see that in the next 2–4 years — your costs for internal controls and independent oversight will go up.
- When models are held accountable for their output, litigation costs for defending rogue model behavior due to negligence will go up.
As each of these scenarios unfold, AI companies will do what all entities do when facing increased expenditures — they will pass those costs along to consumers in the form of higher subscription fees and increased billing on token usage.
What happens when your firm depends on it?
So, when the cost for the plug-ins and prompts go up, what will happen to the advisory firms who have woven the work of their back-office so tightly to their LLM prompts, agentic skills, and AI-generated content that they can no longer unwind from it all?
You can't simply migrate these assets to a new provider the way you can move data. Prompts generate different responses when fed into different models.
Will these advisory firms simply absorb the cost increase? Throttle back on usage? Raise their fees? Or will they be replaced by other firms who chose to incorporate AI in a more discriminate manner, rather than relying on a single technology to run the show?
Use AI without depending on it for everything
I'm not anti-AI, but I'm definitely pro-DPAYEOB (Don't Put All Your Eggs in One Basket).
Use AI where existing tools don't work well. Use AI where it saves the significant time. Use AI where it doesn't put client data, or your reputation, at risk. But don't build your internal processes in such a manner that you can't function without it.
<<Climbs down off soapbox and prepares for the pushback>>