Anthropic launched something useful, without much noise, in May 2026. Claude for Small Business came with 15 ready-made agentic workflows that connect tools most small advice firms already pay for: QuickBooks, HubSpot, DocuSign, Google Workspace, Microsoft 365, Canva and PayPal. There’s no custom build and no API setup. You connect your accounts and the workflows run.

For a firm with one to ten advisers, that matters more than it sounds. Most AI products aimed at financial services are either priced for enterprises or need technical staff that small IFA and wealth management businesses don’t have. This one is built differently, even if the risks are much the same, so here’s what it can and can’t do inside a regulated firm.

What the 15 workflows cover

The workflows sit at Level 1 and Level 2 of the three levels of AI use. Level 1 (education) tasks are things you could already do with a prompt and some patience, now packaged so you don’t have to write the prompt. Level 2 (integration) tasks have Claude working between two systems you already use, passing information across or drafting things that would otherwise be done by hand.

For a wealth management firm, that looks like this:

  • CRM follow-ups. Connected to HubSpot, Claude can draft a personalised follow-up after a meeting note is logged, using context from the CRM record.
  • Document preparation. Connected to Google Workspace or Microsoft 365, it can draft template letters, meeting summaries or review packs from a brief you give it.
  • Contract and form routing. Connected to DocuSign, it can fill in standard documents and send them for signature without anyone re-typing the data.
  • Bookkeeping questions. Connected to QuickBooks, it can answer questions about the P&L, produce summaries or flag anomalies for a person to look at.

None of these replaces a regulated decision. All of them reduce the admin around regulated decisions. Under UK rules, that distinction matters a great deal.

Where the FCA limits sit

The FCA’s existing rules apply in full, and a 2026 joint statement from the FCA, the Bank of England and HM Treasury on frontier AI and cyber risk signalled that AI governance for UK financial services firms is moving from internal good practice towards a firmer regulatory expectation[1]. Everything below follows from that.

Running any AI assistant in a regulated advice firm is now a real operational risk. It goes beyond staff misusing it. It includes prompt injection, people manipulating a chatbot’s persona, and attempts to talk it past its guardrails. Those risks apply to any firm using a connected AI workflow, Claude for Small Business included, and not only to large banks. Treat these tools as operational infrastructure, because that’s increasingly how the FCA sees them.

In practice, three things follow:

  • Claude can support these processes. It can’t own them. Any output that touches a regulated decision, a suitability judgement, a client communication about investments or an AML-relevant check has to be reviewed and signed off by someone with the right permissions. The workflow drafts and the adviser approves. That’s the correct model, not a limitation to work around.
  • Consumer Duty raises the bar for client communications. Under Consumer Duty, communications have to support good client outcomes[2]. An AI-drafted message that hasn’t been reviewed by someone who knows the client doesn’t meet that bar by default. A workflow that drafts, with a person who reviews and personalises before sending, is a reasonable model. A workflow that drafts and sends isn’t.
  • Get your data in order before you connect anything. Connecting HubSpot, DocuSign and Claude means client data moves between those systems. Make sure what you hold is within your consent and retention framework before you wire them together. If your CRM holds data you haven’t audited recently, fix that first.

How to start using it safely

Four steps, in this order.

Pick the workflows you need

Go through the 15 and find the three or four that fix a real problem in how you work today. For a small advice firm, the Google Workspace document drafting and the HubSpot follow-ups are likely to be the most useful straight away. Start with one.

Define the human review step first

For each workflow, write down what the review step looks like, who does it and what they’re checking for. If you can’t describe that step, the workflow isn’t ready to run. That written step is what makes the tool reduce risk instead of adding it.

Check the commercial agreement

Before your firm connects client data to any AI system, your review of the terms should cover where data is processed and stored, what Anthropic can do with your inputs, and how disputes about outputs would be handled. If your usual IT procurement checks haven’t been updated for AI vendors, now’s the time. The guide to reading an AI vendor contract lists the clauses to look for.

Write down how you use it

Firms that can show they thought carefully about AI use, defined their oversight steps and applied them consistently are in a much better position than firms that can’t. A short internal note covering what you use, why, and how you check the output is worth writing now, before anyone asks for it.

Vendor risk

Any AI product built on top of someone else’s model carries a continuity risk. The AI industry is under heavy margin pressure, and product lines aimed at small businesses can be discontinued, repriced or moved upmarket at short notice. If a workflow becomes part of how you operate and the vendor changes direction, that’s a business continuity problem.

That isn’t a reason to avoid the tools. It’s a reason to use them for tasks you could take back to a manual process, or move to an alternative, without a crisis.

Is Claude for Small Business worth it for an advice firm?

For the admin, yes. Ready-made workflows connected to tools a small advice firm already uses mean Level 2 integration work that used to need a consultant or a technical hire is now within reach of an owner willing to spend an afternoon setting it up.

The regulatory constraints don’t change. Output review, data governance and recorded oversight aren’t optional for an FCA-regulated firm. Within those constraints, the tools are useful for taking admin off the functions that don’t need regulated judgement.

This article is for information only. It isn’t regulated financial advice, legal advice or a compliance opinion. Data protection obligations depend on your circumstances and jurisdiction, so consult a qualified compliance professional, solicitor or data protection adviser for advice specific to your firm.

Sources

[1] FCA, Bank of England, and HM Treasury, Joint Statement on Frontier AI and Systemic Cyber Risk, 2026. Source for the regulatory direction on AI governance for UK financial services firms.

[2] FCA, Consumer Duty (PS22/9), July 2022. Source for the requirements on client communications to support good outcomes, https://www.fca.org.uk/publications/policy-statements/ps22-9-new-consumer-duty