1. The CRM is dissolving into other interfaces
A cluster of announcements from Salesforce, HubSpot, Zeotap, and Zendesk, reported by CX Today on September 21, 2026, points to a single directional shift: CRM data is being pushed out of the traditional CRM interface and into AI agents, collaboration tools, and external workflows.
Salesforce used Dreamforce 2026 to introduce AIforce, described by CX Today as "its umbrella term for agentic experiences that bring Salesforce context, permissions, workflows, and actions into external interfaces such as Slack, Microsoft Teams, and Anthropic's Claude." The company also announced ClaudeForce (access to Salesforce data from within Claude Enterprise) and Slackforce (live reports and dashboards built from prompts inside Slack). Parker Harris, Salesforce co-founder, stated: "You may never log in to Salesforce again."
HubSpot, at its Fall 2026 Spotlight, took a different path: a self-updating Smart CRM that automatically captures calls, emails, and meetings, paired with a Context Score that flags where customer data is incomplete. Duncan Lennox, HubSpot's Chief Product and Technology Officer, said: "We hear from customers every day that they don't want to think about which AI tools to use, they just want outcomes."
Zendesk introduced Specialized AI Agents designed to act across connected systems. Zeotap launched an Open CDP built on Apache Spark and customer-owned lakehouse storage.
As CX Today summarized: "The real battle is over who can make that information useful, current, and safe enough for AI to act on it."
"You may never log in to Salesforce again."
2. Why email and campaign teams should pay close attention
These announcements are framed around CRM strategy, but their implications land directly on campaign operations. Every triggered email, every nurture sequence, every personalized send depends on the quality and accessibility of the contact record sitting underneath it. When CRM data begins flowing through multiple new interfaces (AI agents, Slack, Claude), the surface area for both enrichment and corruption expands.
HubSpot's self-updating CRM is a good illustration. Automatic capture of calls, emails, and meetings could mean richer behavioral signals feeding multi-touch campaigns and scoring models. But CX Today raised the obvious counterpoint: "Automatic capture is helpful; accurate customer context is the harder job." A contact record polluted by misattributed meeting notes or duplicate call logs will degrade, not improve, segmentation and personalization.
Salesforce's approach introduces a different risk. If CRM records can be updated from within Slack or Claude, campaign teams lose a single audit trail. CX Today flagged this directly: "Leaders will need to know exactly which customer data an AI agent can see, what it can change, and how those changes are tracked." For any organization running automated always-on campaigns, a field change made by an AI agent in Slack could trigger or suppress a send that no human reviewed.
This connects to a pattern we have explored before: the data trust crisis that comes from operational neglect.
"We hear from customers every day that they don’t want to think about which AI tools to use, they just want outcomes."
3. Governance is now a campaign deliverability problem
When CRM records were updated by humans inside a single application, governance was a matter of field-level permissions and validation rules. When those same records can be modified from Claude, Slack, or an AI agent inside Zendesk, campaign operations teams inherit a governance challenge that traditional marketing automation platforms were not designed to address.
CX Today noted that as CRM actions move into new interfaces, "the governance burden gets heavier." That burden falls on data management and campaign reporting teams in equal measure.
Consider auditing every field your marketing automation platform reads from CRM before enabling agentic write paths. CX Today's framing is worth adopting as a test: can you answer which customer data an AI agent can see, what it can change, and how those changes are tracked? If not, the governance model needs work before a new interface is enabled.
For organizations running on HubSpot, the new Context Score could become a useful diagnostic if connected to campaign workflows. We recommend evaluating whether a low Context Score for a segment about to enter a nurture sequence should pause or flag that send, rather than treating it as a CRM hygiene metric alone.
4. Practical steps for campaign operations teams
Consider running a campaign maturity assessment with a specific focus on how CRM data reaches your marketing automation platform. These vendor announcements multiply the number of systems that can write to the contact record, and campaign teams need visibility into those write paths.
We recommend that teams using Salesforce evaluate ClaudeForce and Slackforce access controls before those features reach production. CX Today reported that Salesforce is "positioning itself as the governed system underneath AI interactions," but governance configuration is the customer's responsibility. Test against a scenario where an AI agent writes an incorrect value to a field that drives email personalization or send suppression. What happens to your next scheduled campaign?
For HubSpot users, the self-updating Smart CRM creates an opportunity to improve lead nurturing signal quality. But as CX Today noted, the challenge is "whether HubSpot can make 'touchless' CRM data reliable enough for customers to trust it." Duplicate activity logging, for instance, could inflate engagement scores and push contacts into sales-ready sequences prematurely.
As we noted in our analysis of marketing automation platforms being rebuilt around context, the shift from campaign-centric to context-centric architectures is real. These CRM announcements accelerate it. Campaign teams that treat this as someone else's infrastructure problem will find their sends increasingly driven by data they did not verify, written by agents they did not configure.
