Claude for advisors hits the workflow wall

Anthropic's Claude for Financial Advisors arrived with a statistic attached: the Kitces research it cited holds that a typical advisory practice spends about a sixth of its time in client meetings, a number that has held steady for years. The optimism in the launch coverage was genuine, but nearly every analyst who supplied it attached a condition, because the consultants who live inside advisor tech stacks understand that moving that sixth depends on workflow at least as much as on model quality.

Alois Pirker, the longtime analyst who now runs his own consultancy, Pirker Partners, called the outreach ahead of the launch a good sign and then named the wall he expects Claude to hit. Salesforce for Financial Services reached general availability a decade ago, later becoming Financial Services Cloud, and the firms needing real customization landed on Practifi, which is built on top of it; horizontal platforms stall, in his read, at the point where a practice wants a process that does not exist in the box.

His illustration travels well into a partner meeting. LPL Financial runs more than 30,000 advisors, and Pirker reckons that single firm holds 15 different types of advisors, each expecting something specific from the software. A model tuned for the median practice is tuned for no practice in particular, and the firm paying for it is the one that has to decide which of the 15 gets served first.

Will Trout, who directs securities and investments research at Datos Insights, moved the test to integration. Many vendors have promised workflow efficiency, he wrote, and the connectors to Schwab, Addepar, and Orion are the proof of concept; if integration works, he said, this becomes a margin story. The practical bar is that the model's output has to terminate inside the systems a practice already runs, or the practice runs the work twice.

The Schwab-Claude integration this publication priced at $240 a year per seat landed on roughly 16,000 RIAs, and the license was never the expense. The expense is process work: choosing which practice types get which workflow, naming who owns the output, and knowing what the compliance file looks like when a model drafts client correspondence. Connectors get a firm only as far as a demonstration.

So the judgment for the next few quarters is that seat count will separate almost nothing. The practices that gain will be the ones that pick two workflows, write down how long each takes today, and can produce the delta a year out. Claude will likely be good enough to move that number. Firms that skipped the baseline will have no way to tell whether it did, which is a harder thing to explain to a recruiting candidate than the price of a license.