Verdict

“The tool's effectiveness relies on correctly combining data from multiple undocumented API endpoints to get a comprehensive view of user activity.”

The idea is to build a tool that helps Claude Teams admins identify dormant seats by leveraging undocumented internal API endpoints. The concept is feasible for a solo or 2-person team to build in 4-12 weeks. The provided Python code serves as a good starting point, and the required technical expertise is relatively low. However, there are some caveats to consider: the undocumented endpoints may change, and the sessionKey cookie is sensitive information. Additionally, the rankings endpoint is capped at top 100, which limits the accuracy of identifying dormant seats. Despite these limitations, the tool can still provide value by narrowing down the candidate set. The biggest challenge lies in handling the quirks of the undocumented endpoints and ensuring the tool remains functional if Anthropic changes the API. Overall, the technical complexity is moderate, and the timeframe is realistic for a small team.

“Only by scraping undocumented internal APIs can an entrant reveal true per‑user activity and differentiate from existing admin tools, but this creates a fragile, non‑scalable advantage.”

The idea solves a genuine pain point for Claude Teams administrators by exposing activity data that the official UI and documented APIs deliberately hide. Existing competitors—such as generic SaaS admin consoles (e.g., Slack, Microsoft Teams) and Anthropic's own limited Admin API—do not provide per‑user last‑active or Claude Code usage metrics, leaving admins to rely on aggregate dashboards or manual outreach. By reverse‑engineering the undocumented internal endpoints, the entrant can compile a unified view of chat and code activity, identify dormant seats, and enable targeted coaching or seat reclamation. This differentiation is real because no current solution offers a comparable, out‑of‑the‑box export of granular, cross‑product activity. However, durability is weak: the reliance on undocumented, session‑based endpoints makes the approach fragile; any change to API paths, authentication, or rate limits could break the script, and Anthropic could withdraw access without notice. Moreover, the solution is a DIY script rather than a polished product, limiting its market appeal and defensibility. While the niche need is clear and the method currently works, the lack of a sustainable moat and the high maintenance burden reduce the long‑term defensibility, warranting a mid‑range score rather than a high one.

“Reliance on undocumented, unstable APIs and manual workarounds severely limits scalability and longevity.”

The venture's viability is severely threatened by its reliance on undocumented, ephemeral endpoints, coupled with the lack of a scalable, authorized data access method. Regulatory or platform changes could immediately render the solution obsolete. Furthermore, the need for manual intervention (e.g., sessionKey handling, post-script investigation of 'dormant candidates') limits scalability and introduces significant operational overhead, deterring broad adoption. The niche appeal (only to Claude Teams admins with specific pain points) also constrains the customer base, exacerbating churn risks if the workaround is invalidated.

“Claude Teams admins are flying blind on seat utilization because Anthropic hides user-level activity data, forcing them to reverse-engineer internal APIs — a clear signal of an unmet enterprise need with real budget at stake.”

This idea targets a real, urgent, and unmet need among Claude Teams administrators managing 50+ seats — a segment likely numbering in the tens of thousands globally, given Anthropic’s enterprise growth. These admins are financially responsible for seat utilization, yet Anthropic provides zero visibility into per-user activity beyond aggregated dashboards. The 30% gap between SSO logins and claimed MAU represents real wasted budget — seats that could be reclaimed, reassigned, or used for coaching. The fact that admins are reverse-engineering undocumented APIs proves demand is strong and underserved. The gotcha about Claude Code being siloed from chat activity is particularly valuable: it reveals a systemic blind spot that even power users miss, making this solution not just useful but necessary for accurate auditing. While the method is brittle (relying on cookies and undocumented endpoints), it’s a stopgap for a product gap that Anthropic has ignored for months. This isn’t a niche hack — it’s a critical operational tool for any org spending $10k+/year on Claude Teams. The audience is enterprise IT, procurement, and internal AI adoption leads who care deeply about ROI and compliance. They have budget (Claude Teams starts at $30/user/month), and they’re actively seeking solutions. The fact that the author had to build this themselves, and that no official API exists, confirms market failure. This idea has high perceived value, low friction to adopt (Python script), and massive upside if Anthropic eventually adopts it as a feature. The only reason it’s not a 10 is the fragility of the implementation — but that’s a product problem, not a market one.

“This exploits a high-value gap in Anthropic’s admin tooling with a near-zero-cost solution that directly ties to measurable cost savings.”

This idea addresses a critical, unmet need for Claude Teams admins: identifying dormant seats to reclaim costs or coach users. The revenue model is indirect but powerful—it enables cost savings (reclaiming unused seats at ~$20-30/user/month) or efficiency gains (coaching inactive users to adopt tools). The 'product' is a script/tool (or future SaaS) that automates the undocumented API workflow, with pricing potential as a one-time license ($50-200) or subscription ($10-50/month) for enterprises. Unit economics are strong: near-zero marginal cost (API calls are lightweight), and the value captured (10-30% seat reclamation) far exceeds the price. Channels include direct outreach to Anthropic admins (Slack/Discord communities, LinkedIn), GitHub for the open-source script, or a paid wrapper with UI/alerts. Gross margins would be ~90%+ for a SaaS version. The main risk is Anthropic shutting down the undocumented endpoints, but the urgency of the problem (no official solution) and the viral potential among admins mitigate this.