BlockSentient Independent SaaS audits

AI Visibility Report: Agentforce Service

Public product evidence and observed AI-agent activity for Agentforce Service.

1 evidence issue found · 10 buyer questions checked · Checked 14 Aug 2026

Buyer question

Can an Enterprise customer deploy Agentforce Service Agent, and which entitlement is required?

CONFLICT High buyer impact

Evidence

The Service Agent row lists Enterprise, Performance, Unlimited, and Developer editions with Foundations or Agentforce 1 Editions, while noting that some standard agent actions require additional add-on licenses.

Agent Types and Considerations

View source
The Compare Availability and Billing table lists Agentforce Service Agent for Performance, Unlimited, and Developer editions with Einstein for Service, Einstein Platform, Agentforce Service Agent, or Einstein 1 add-ons, omitting Enterprise.

Compare Einstein Bots with Agentforce Service Agent

View source

What we found

Salesforce pages give different edition requirements for Agentforce Service Agent

Why it matters

One current Salesforce page includes Enterprise for Agentforce Service Agent, while another current availability table omits Enterprise and lists different qualifying add-ons.

Suggested fix

Publish one Agentforce Service Agent edition and entitlement matrix and align the overlapping Help pages.

Evidence observed 14 Aug 2026

Complete evidence check · EUR 19

Up to three evidence issues, exact first-party sources, and suggested fixes.

Request the complete check

Early access. No payment today.

Supporting AI-agent activity

Reporting period: Aug 10 to Aug 16, 2026 (Mon to Sun, 7 days)

5 recognized AI-agent requests
2 distinct agents
0% live retrieval
100% training/index

Daily activity

Daily recognized AI-agent requests
Monday 10 August 20260
Tuesday 11 August 20260
Wednesday 12 August 20260
Thursday 13 August 20260
Friday 14 August 20261
Saturday 15 August 20262
Sunday 16 August 20262

Agent breakdown

Agent Operator Mode Requests Share
Meta-ExternalAgent Meta Training/index 4 80.0%
GPTBot OpenAI Training/index 1 20.0%

Methodology and limitations

Evidence check: We evaluate a defined sample of current public first-party product documentation against buyer-critical questions. Evidence states describe what the public sources support at the time of the scan.

AI-agent activity: Live retrieval identifies recognized agent identities associated with on-demand retrieval. Training/index identifies crawler activity associated with indexing or model-related collection.

Important limitation: A recognized AI-agent request does not prove that the page was used in an answer, cited, recommended, or triggered by a unique human user.

Scope: This report evaluates public documentation. It is not a product-quality rating, compliance certification, security assessment, legal conclusion, or statement about private vendor systems.

Refresh: Evidence integrity is refreshed monthly when a current evidence record is available. AI-agent activity statistics refresh weekly.