Sites Adoption Stalls Before First Creation
OpenAI’s ChatGPT Work demo argues that product teams should not treat an adoption-funnel drop-off as its own diagnosis. Using synthetic data for a product called Sites, it identifies the main loss before users create for the first time, then draws on Slack and Google Drive feedback to attribute the problem to uncertainty about the product’s use cases rather than post-creation retention. The proposed response is to build task-specific templates, target promising enterprise accounts and turn the analysis into an assigned Slack action plan for review.
The bottleneck is before first creation, not after it
Leadership’s question is where to invest in a newly launched product called Sites: interactive web apps that colleagues can create and share. The immediate concern is weak external-customer adoption. The analysis begins by asking ChatGPT Work to assess whether Sites is resonating and where the team should focus, using a Data agent and a Sites Semantic Layer.
The resulting report frames the question as a creation funnel. In the illustrative snapshot, 10,000 people have access to Sites, but only 1,567 start creating. That is 15.7% of those with access. Once people begin, the funnel is comparatively healthy: 1,534 people successfully create a Site, or 97.9% of people who started, and 1,309 publish, or 85.3% of successful creators.
The implication is not simply that Sites fails to retain users. The report’s executive summary says growth is now coming mostly from people who have created before, while the largest opportunity lies in getting more eligible users to make their first Site. In other words, the main loss occurs at the transition from access to action.
That finding narrows where to investigate, but it does not answer why people are failing to start. The demo explicitly distinguishes between two possible explanations: customers may be unable to find Sites, or they may not understand what to use it for. Those diagnoses would imply different product decisions. A discoverability problem could call for changes to navigation or access; an unclear value proposition would call for better positioning, examples, or guided starting points.
Usage data identifies the problem; customer input supplies the diagnosis
The next prompt asks for customer feedback from Slack and Google Drive on why eligible users are not starting Site creation. The purpose is to “double-click” on the funnel drop-off rather than treat a behavioral metric as a sufficient explanation.
The synthesized customer-input themes point toward uncertainty over the product’s role. Participants with access reportedly did not know where to start or whether Sites was intended for public websites, dashboards, or internal team work. The feedback also emphasizes demand for task-based examples: customers want to see how Sites could relate to their own work.
That shifts the interpretation of the funnel. The evidence shown in the demo supports an orientation and positioning problem more directly than it supports a claim that people cannot locate the product. Customers reach the point of eligibility, but many do not have a concrete enough picture of the job Sites is meant to do or the first step they should take.
The data is expressly illustrative and provided for demonstration purposes. Its role is to show a workflow: product-usage data identifies the stage at which adoption breaks down, and customer feedback helps distinguish between plausible explanations for that break.
The proposed response pairs product scaffolding with targeted go-to-market work
On the basis of the customer feedback, the recommendation is to pursue two actions.
First, the product team should build templates around high-value use cases. The intent is not merely to add reusable assets, but to help prospective creators imagine what they could build. Templates turn a general-purpose product into visible, task-specific starting points for customers who are unsure whether Sites fits their work.
Second, the go-to-market team should plan an enterprise campaign for promising accounts identified in the analysis. The stated aim is to enable those customers to see the potential of the product, rather than leaving them to infer its value from access alone.
The demo then turns the analysis into a draft Slack update for the #demo-sites-team channel. The request asks for a concise summary of the adoption analysis, the two action items, and specific ownership. The draft characterizes the central finding plainly: “Sites resonates once customers start — but too few reach first creation.” It assigns a product-side direct responsible individual to the Sites Activation PM partner and a GTM-side owner to the Enterprise PMM lead.
The workflow leaves room for review. The generated Slack post is presented as a draft that can be edited before it is posted. Once reviewed, it is sent to the team as a shared statement of the problem, the proposed response, and accountable owners.