Forrester says clients are now using its Forrester AI tool more often than conventional search to navigate the firm’s research library, a notable behavioral shift for a company whose business has long depended on searchable analyst reports. The metric emerged during the company’s July 30 second-quarter earnings call, as reported by AI Media House, alongside Forrester’s effort to push its research directly into Microsoft Teams and Microsoft 365 Copilot.
For Windows-centric enterprises, the immediate relevance is less about Forrester’s financial performance than the delivery model: subscription research is becoming an in-workflow Copilot experience rather than another web portal employees must visit and search manually.
Forrester reported second-quarter revenue of $100.2 million, down from $111.7 million a year earlier, while contract value fell 3% to $283.2 million. CEO George Colony nevertheless said adoption of Forrester AI is accelerating, and the company plans further AI products and capabilities in the second half of 2026.
Forrester launched its AI agent for Microsoft 365 Copilot on April 28, following a March release of the Forrester AI app for Microsoft Teams. Eligible Forrester license holders can query the firm’s proprietary research and frameworks within Microsoft’s productivity environment instead of switching to Forrester’s own interface.
The company says the agent can provide research-backed answers, summaries and access to source analysis, while its Model Context Protocol connector is intended to let customers bring Forrester material into their own systems. AI Media House reported that hundreds of accounts have been activated for Teams and Copilot access, with Gemini, Claude and Slack integrations in development.
That makes the Microsoft integration strategically important. A knowledge product embedded in the tools where users write documents, prepare presentations and hold meetings has a far clearer path to habitual use than one accessed through a separate search box.
The new internal usage metric gives that theory an unusually concrete demonstration. If Forrester’s clients are choosing a conversational, source-grounded AI tool over indexed search inside the company’s own research environment, it suggests the interface shift is not merely a prediction about public search traffic.
There is an important distinction, however: Forrester AI is a closed, licensed research system, not a broad consumer search engine. Its advantage is that responses are grounded in material customers already pay to access, with links back to the originating research. That is precisely the model enterprise IT teams should favor when evaluating AI knowledge tools: bounded data, permission-aware access and a practical way to verify the answer.
The larger signal is that Microsoft 365 Copilot is increasingly positioned as a workfront for specialist agents rather than a single assistant with generic knowledge. Forrester’s test will be whether easier access improves research subscription retention and contract-value growth while its broader revenue continues to decline.
For IT leaders, the practical question is no longer whether employees will ask AI for answers instead of searching. Forrester’s own data suggests they already are; the next decision is whether the answers come from governed enterprise sources or from the open web.
Forrester reported second-quarter revenue of $100.2 million, down from $111.7 million a year earlier, while contract value fell 3% to $283.2 million. CEO George Colony nevertheless said adoption of Forrester AI is accelerating, and the company plans further AI products and capabilities in the second half of 2026.
Copilot becomes the front door to analyst research
Forrester launched its AI agent for Microsoft 365 Copilot on April 28, following a March release of the Forrester AI app for Microsoft Teams. Eligible Forrester license holders can query the firm’s proprietary research and frameworks within Microsoft’s productivity environment instead of switching to Forrester’s own interface.The company says the agent can provide research-backed answers, summaries and access to source analysis, while its Model Context Protocol connector is intended to let customers bring Forrester material into their own systems. AI Media House reported that hundreds of accounts have been activated for Teams and Copilot access, with Gemini, Claude and Slack integrations in development.
That makes the Microsoft integration strategically important. A knowledge product embedded in the tools where users write documents, prepare presentations and hold meetings has a far clearer path to habitual use than one accessed through a separate search box.
The “visibility vacuum” reaches Forrester itself
Colony has used the term visibility vacuum to describe a B2B buying environment where buyers conduct more research through AI answer engines and leave fewer of the trackable search, web and content signals that sellers have traditionally relied on. Forrester’s B2B Summit North America agenda in Phoenix prominently framed the challenge as “Overcoming The Visibility Vacuum.”The new internal usage metric gives that theory an unusually concrete demonstration. If Forrester’s clients are choosing a conversational, source-grounded AI tool over indexed search inside the company’s own research environment, it suggests the interface shift is not merely a prediction about public search traffic.
There is an important distinction, however: Forrester AI is a closed, licensed research system, not a broad consumer search engine. Its advantage is that responses are grounded in material customers already pay to access, with links back to the originating research. That is precisely the model enterprise IT teams should favor when evaluating AI knowledge tools: bounded data, permission-aware access and a practical way to verify the answer.
The enterprise consequence is governance, not novelty
Forrester has said its Teams app does not use submitted prompts or generated responses to train models, a policy that addresses a central concern for organizations placing proprietary work context into an AI interface. Administrators still need to confirm identity controls, licensing, retention, auditability and the scope of content the agent can retrieve before presenting such tools as a replacement for internal search.The larger signal is that Microsoft 365 Copilot is increasingly positioned as a workfront for specialist agents rather than a single assistant with generic knowledge. Forrester’s test will be whether easier access improves research subscription retention and contract-value growth while its broader revenue continues to decline.
For IT leaders, the practical question is no longer whether employees will ask AI for answers instead of searching. Forrester’s own data suggests they already are; the next decision is whether the answers come from governed enterprise sources or from the open web.
References
- Primary source: AIM Media House
Published: 2026-07-31T10:55:24.062000+00:00
AIM Media House - AI, Technology & Business Insights
aimmediahouse.com
