Salesforce and agentic AI for insurance: investing in foundations

Jul 01, 2026 AI in Insurance, Core Systems, Salesforce

Salesforce remains one of the most important CRM systems used by insurers across many markets. With the introduction of “Headless 360”, Salesforce enables insurers to build an infrastructure for agent-based process design and AI-supported automation. This approach allows companies to customize more easily and adapt processes to specific requirements. As a result, effort is reduced, and implementation takes less time. However, insurers that want to adopt this approach must first prepare their infrastructure accordingly.

A platform in transition

Something has changed in how Salesforce positions itself. For years, the implicit message was: build everything here, make Salesforce the center. This perspective is increasingly fading. The platform now sees itself more as a building block within a larger architecture, not the architecture itself.

For those who previously found the old approach discouraging, this is a significant shift. The new “Headless 360” solution expresses this most clearly: every Salesforce function – workflows, data, business logic, agent actions – is now made available as an API or as a tool within the Model Context Protocol (MCP). MCPs allow LLM models to access different data sources and tools in a standardized way.

With “Headless 360,” Salesforce provides core platform functions. Data, workflows, and business logic are directly accessible through more than 60 MCP tools, as well as APIs. At the same time, the provider is expanding its development environment with “Agentforce Vibes 2.0,” adding AI-supported capabilities.

With Headless 360, Salesforce shifts access away from the graphical user interface toward APIs. Applications are no longer used exclusively via user interface (UI) interactions but are accessed through APIs, MCP tools, and automated agents. Salesforce has been designed for human users. With Headless 360 it becomes a system for AI Agents.

Why insurance is a complicated case

CRM has never played as central a role in insurance companies as it has in other industries such as banking or retail. Customer relationships in insurance are often indirect - via brokers or comparison platforms - so building a strong CRM layer was not a top priority for a long time. This history is important when evaluating Headless 360.

For insurers that are now seriously investing in an IT infrastructure with “Customer 360” capabilities, the traditional Salesforce user interface is usually sufficient. There is no strict need to adopt Headless 360. The standard solution is already powerful. Additional architectural complexity without clear benefits seems unnecessary.

However, there is a more interesting perspective: Salesforce positions the messaging service Slack – and alternatively Microsoft Teams – as a conversational interface for CRM. For teams working on complex cases, such as underwriting large risks or handling disputed claims, this is a practical entry point. CRM data becomes visible directly within conversations, without needing to open a separate system. This represents a gentler entry point and is particularly important for organizations that have previously used CRM more cautiously.

The unpopular topic of business rules

A common assumption about agent-based platforms is that AI agents can simply take over existing business rules. This may hold true for newly built system landscapes. In most legacy insurance systems, however, reality is often more chaotic: rules are distributed across configuration files, application code, and sometimes even the front end. They are rarely structured clearly or organized in a service-oriented way.

A realistic approach is to have agents generate suggestions while the domain system performs the actual validation. This is less spectacular than full autonomy, but it honestly reflects the current state in many companies – and builds trust in the technology before further expansion.

The real question is not technical

“Build once, deploy anywhere” is an attractive idea – but difficult to implement. Not because the technology doesn’t work, but because it requires a level of discipline in enterprise architecture and reuse governance that many organizations still need to develop. Headless 360 provides the possibility – but it does not replace this discipline.

It can be seen this way: a tightly integrated platform means many decisions have already been made – consistent user interfaces, defined use cases, controlled connections between tools. A composable, “headless” approach, on the other hand, means that these decisions must be made by the company and its users. This is only an advantage if there is a clear vision of what should be achieved and the governance to implement it consistently. Otherwise, it leads to additional complexity without corresponding benefits.

If a chief technology officer were to ask, “Should we prepare for this?” the answer would be: yes, but probably not with a concrete Headless 360 project. The better investment lies in the foundations: capabilities in enterprise architecture, a clear understanding of one’s own differentiators, and the discipline to reuse components instead of rebuilding them. These factors truly pay off, regardless of how platforms evolve.

The direction Salesforce is taking is understandable. The key question is whether insurers are ready to use it effectively – and that is an internal question, not a platform issue.