Salesforce is turning the Sales Cloud into a background platform. Two new AI agents take over the manual work: “Momentum” captures conversation context automatically, “Hunter” triggers outbound actions on its own. For B2B sales leaders and CRM teams in mid-sized companies that means a pipeline that fills more proactively and less maintenance effort. In return, implementation shifts away from screen layouts towards data quality and guardrails for autonomous AI cadences.
Headless CRM
Salesforce is pushing the headless approach hard. Data, business logic and AI functions detach from the classic Salesforce interface. They move to where your sales team already works: Slack, Microsoft Teams or external AI tools such as Claude from Anthropic. Through the Model Context Protocol (MCP), these external applications access CRM data and actions directly. You can, for example, have Claude produce a pipeline review without leaving the chat. The system pulls open pipelines and deal stages and prepares them for you.
Momentum: CRM context from every conversation
When a prospect raises concerns in a video call or sends new requirements by email, those details are often missing in the system. The new Momentum technology captures such conversations from voice and video applications, text messages and emails. It transforms the unstructured data and writes these “deal insights” into the Salesforce records in real time. If a deal is at risk, management automatically receives warning signals in Slack or Teams. Momentum is expected to reach general availability in October 2026.
Hunter: the autonomous outbound agent
The existing agent Piper focuses on inbound leads, Hunter takes over the outbound side. It learns from the success patterns of your best sellers and scores prospects. When a sales rep’s pipeline runs dry, Hunter proactively identifies new contacts. It explains its selection and creates a multi-step outreach cadence of calls, emails and further touchpoints. Hunter is currently in a pilot phase; general availability is planned for November 2026.
Which AI agents exist for inside sales?
In inside sales (SDR) and sales support, AI agents mainly take on these tasks:
- Lead qualification: agents analyse incoming requests from forms or emails in real time. They score leads by criteria such as company size or budget and enrich profiles with external data. Warm leads go to sales, cold ones into a nurturing flow.
- First contact (SDR): these agents greet prospects, ask qualifying questions, check the fit with the ideal customer profile (ICP) and, if needed, book an appointment directly in the responsible employee’s calendar.
- CRM maintenance: they document calls or email threads promptly and in a structured way. The agents create conversation summaries and update deal stages in the CRM fully automatically.
Prerequisites for autonomous agents in mid-sized companies
When Hunter sends emails on its own or Momentum transcribes internal meetings, technical and legal foundations come into focus. Before such agents go into production, you have to clarify three areas:
- If the data foundation is wrong, an agent draws false conclusions or exposes sensitive deal information to unauthorised people. Check outdated contacts and incorrectly set permissions.
- The automatic recording of video calls and emails by Momentum touches the GDPR and the personal rights of employees. Involve the works council early and establish clear opt-in processes for customers.
- An autonomous agent such as Hunter needs precise guardrails. Define exactly which tone it uses, how often it follows up and at which signals it hands over to a human colleague. Otherwise AI-generated messages damage the customer relationship.
As a first step, review your current permission structures in Salesforce and clean up outdated pipeline data before you activate autonomous cadences. Want to evaluate whether your environment is ready for Momentum and Hunter? Let us look at your data quality together. Get in touch to discuss the next steps for your B2B sales.
