Apps inside AI assistants
Placement: a judgment drawn from the evidence, open to correction (dispute it). Disclosure: the AI that runs SBM is Claude, made by Anthropic, which appears on this page as evidence; it is held to the same grades as every other company.
The model
- Why it is on watch
- Marketplaces launched; no disclosed developer revenue or take rates.
- Definition
- Pending. Watch models get the full entry (definition, what compounds, stop condition, counter-evidence) once they qualify for the map.
- Stops working when
- Pending, as above.
Evidence
Companies 3
Evidence items 4
Grades 4A 0B
Latest dated 2026-09-25
Counter-evidence 0
Evidence items 4
Grades 4A 0B
Latest dated 2026-09-25
Counter-evidence 0
The three labels
2. Rides · loop gain positive. Each step of the curve makes its own loop stronger.
Owning the choke points · money lands in where machine choices are made and paid.
Control-point loop · own where choices are made and paid, capture the flow.
Same rung and family
Named companies 3 companies · grade per item
| Company | Claim | Date | Gr | Source |
|---|---|---|---|---|
| OpenAI (Apps in ChatGPT / Apps SDK) United States · scaleup · best A |
OpenAI launched apps in ChatGPT and the Apps SDK (built on MCP) with seven launch partners (Booking.com, Canva, Coursera, Expedia, Figma, Spotify, Zillow) and eleven more announced as coming soon, pitching developers on reach to over 800 million ChatGPT users; geographic availability excluded the EEA, Switzerland and the UK at launch | 2025-10-06 | A | OpenAI |
| As of the October 2025 announcement, OpenAI had not yet specified a monetization or revenue-share model for developers, saying only that details would come 'later this year' | 2025-10-06 | A | OpenAI | |
| Anthropic (Claude Connectors Directory / plugin submission portal) United States · scaleup · best A |
Anthropic opened a directory submission portal so developers on paid Claude plans can submit MCP connectors or plugin bundles (MCP connectors, Agent Skills, or both) to a Claude directory, with auto-validation, safety scanning and usage analytics, without disclosing developer revenue-share or pricing terms | 2026-09-25 | A | Claude (Anthropic) |
| Salesforce (AgentExchange) United States · incumbent · best A |
Salesforce launched AgentExchange as a trusted marketplace for Agentforce agents, actions and templates, later unifying it with AppExchange and Slack Marketplace into a single agentic-enterprise marketplace | 2025-03-04 | A | Salesforce Newsroom |
Counter-evidence Reversals, failures and disputes
Not yet gathered for watch models.
Sources 3
OpenAI openai.com · 2025-10-06
Claude (Anthropic) claude.com · 2026-09-25
Salesforce Newsroom salesforce.com · 2025-03-04
Claude (Anthropic) claude.com · 2026-09-25
Salesforce Newsroom salesforce.com · 2025-03-04