Salesforce releases Headless 360, transforming the entire platform into an AI agent infrastructure

robot
Abstract generation in progress

ME News message: On April 17 (UTC+8), according to Beating Monitoring, Salesforce announced Headless 360 at the TDX Developer Conference, opening up the core capabilities of the entire platform as APIs, MCP tools, and CLI commands. This enables AI agents to directly call enterprise data, workflows, and business logic without having to open the Salesforce backend. The first batch of capabilities available today exceeds 100 in total, including more than 60 MCP tools and more than 30 pre-configured development skills, which can connect with external programming agents such as Claude Code and Cursor.

This is more like an architectural pivot than a product update. Salesforce has said that the default main entry point for its enterprise software will shift from its own website to platforms such as Slack, ChatGPT, and Claude; what it needs to retain is the underlying enterprise context, permissions, and processes.

Headless 360 also launched the Experience Layer, rendering the same set of interactive components across different interfaces such as Slack, mobile devices, and ChatGPT; it further added governance tools including Testing Center, Custom Scoring Evals, and A/B Testing.

Salesforce also stated that the DevOps Center MCP can consolidate build processes that previously required switching between four tools into a single experience, potentially shortening the development cycle by up to about 40%.

(Source: BlockBeats)

View Original
This page may contain third-party content, which is provided for information purposes only (not representations/warranties) and should not be considered as an endorsement of its views by Gate, nor as financial or professional advice. See Disclaimer for details.
  • Reward
  • 11
  • 10
  • Share
Comment
Add a comment
Add a comment
SushiRebalance
· 6h ago
A/B testing is built-in directly; for AI product iteration, this is really necessary to avoid setting up a separate system.
View OriginalReply0
TideShellLedger
· 16h ago
The first batch of capabilities is over a hundred items.
If the iteration speed can't keep up, it will quickly become technical debt.
Let's see how they maintain it.
View OriginalReply0
RiskOffRina
· 22h ago
CLI + MCP + API three-layer openness, with a comprehensive developer experience, covering everything from scripts to proxies.
View OriginalReply0
BoringButBullish
· 22h ago
Shorten the development cycle by 40%. Wait for real case studies to verify, just listen to the numbers at the press conference.
View OriginalReply0
PerpNightRunner
· 22h ago
Supporting external proxy access is a good thing, but could it become another entry point for vendor lock-in?
View OriginalReply0
GaslightGardener
· 22h ago
Traditional CRM giants transforming into AI infrastructure providers, headless architecture is indeed a way out.
View OriginalReply0
RollupStreetKid
· 22h ago
DevOps Center MCP integrates multiple tool workflows, reducing development cycles by 40%, which should be quite attractive to engineering teams.
View OriginalReply0
MevBreakRoom
· 22h ago
I'm more curious about how corporate data permissions are managed. With APIs open, where are the security boundaries?
View OriginalReply0
GlassFishTankArbitrage
· 22h ago
Experience Layer and Testing Center these governance tools are keeping up, indicating that they are also enhancing the controllability of AI agents.
View OriginalReply0
QuietExitPlan
· 22h ago
Over 60 MCP tools + 30+ development skills, the first batch opens hundreds of capabilities, this ecosystem is being built quickly.
View OriginalReply0
View More
  • Pinned