Claudeforce: Salesforce moves into Claude — what about your CRM?

On 26 August 2026, Salesforce and Anthropic announced Claudeforce, a partnership that pulls the entire CRM inside Claude: the “Salesforce in Claude” plugin lets a rep prep for a meeting, review deal health and update the pipeline without ever opening Salesforce. For an SME paying per seat, per month for its CRM, the question this reopens isn’t just “what does this plugin do” — it’s the one we field constantly at Evicron, an AI and custom software studio based in Barcelona: does a custom CRM still make sense once the big vendors start living inside the chat you already use all day?

What Claudeforce actually is

Claudeforce is the expanded strategic partnership between the two companies, per Salesforce’s official announcement and CNBC’s coverage. Its first concrete piece is “Salesforce in Claude”, a plugin with 37 prebuilt sales skills — meeting prep, deal health review, pipeline management and updates — that lets a seller reason over live revenue data and take governed action directly from Claude, with Salesforce’s CRM as the source of truth behind it.

This isn’t a chatbot bolted on top of the CRM. The CRM is the one moving to where the rep already works.

How it works under the hood

The plugin runs on AIforce, Salesforce’s enterprise framework that connects business data and workflows to AI agents through MCP servers, APIs and command-line tools — the same agent-connection standard we covered when it spread to other CRM platforms. In practice, an admin connects the plugin once: authentication and permissions apply across the whole team from day one, with no per-user setup.

Timeline and who gets it

Right now, “Salesforce in Claude” is available to a small group of pilot customers, with an open beta expected in September 2026 and more skills planned for later this year. It isn’t something any SME can switch on from its Salesforce dashboard tomorrow, but it signals a direction: major enterprise SaaS players are betting heavily on living inside an AI assistant’s interface rather than the other way round. Trade press has already framed it as a response to “SaaSpocalypse” worries — the concern that AI chat interfaces could make traditional SaaS apps irrelevant.

An important distinction: this isn’t a customer-facing agent

Don’t confuse Claudeforce with the conversational agents that talk directly to an end customer, which we already covered when looking at HubSpot’s Agent Hub and Salesforce’s Agentforce: those tools fall squarely under the transparency obligations of Article 50 of the EU AI Act, because they interact with people outside the company. “Salesforce in Claude” is, for now, an internal tool for your own sales team to work with its own data from inside Claude: it doesn’t by itself change what your company must disclose to a customer, though it does change where and how your team works with that data.

Does this make a custom CRM pointless?

No — for the same reason we laid out when comparing a custom CRM to Salesforce or HubSpot: Claudeforce makes it more comfortable to work inside Salesforce’s standard mould, but it’s still that mould. If your sales process fits the fields and flows Salesforce already offers, this is genuinely good news: fewer clicks, fewer open tabs, an assistant that already knows how to query your pipeline. But if your competitive edge is precisely the process no off-the-shelf CRM models well — like the custom mortgage CRM we built for Hipteca, with OCR for handwritten forms and an assistant querying data no standard Salesforce field accounts for — a plugin that makes the standard CRM smoother doesn’t solve that problem: the missing custom data model is still missing.

What Claudeforce does raise is the bar for comparison: if you’re sticking with a standard CRM, demand that its AI layer works this well integrated. And if you’re building one from scratch, it’s worth designing in a conversational assistant tied to your own data from day one, not bolted on as an afterthought.

What to ask before jumping on this

1. Does the agent only touch internal data, or does it talk to customers too? The answer changes whether AI Act transparency obligations kick in.
2. What happens to your data if you switch AI vendors? A plugin connected via MCP to a standard CRM is still bound by that CRM’s terms.
3. Does the cost scale with usage, or is it a flat licence? Current pilots don’t publish pricing; project it out before committing once it leaves beta.
4. Does your actual process fit the standard CRM, with or without this plugin? If the answer is still no, the plugin doesn’t change the underlying equation.

How we see it at Evicron

At Evicron we’ve spent since 2019 helping companies make this call without a conflict of interest: sometimes the honest recommendation is to make good use of your standard CRM’s new AI features, and sometimes it’s to build something of your own because the process you want to automate doesn’t fit any off-the-shelf mould. In our AI consultancy for businesses we start by auditing what your actual sales process needs before recommending a platform or a custom build.

Want us to work out whether it pays to lean on your current CRM’s new features or build one of your own? Get in touch: the first discovery session is free and we reply within 24 hours.