Goldman Sachs: 26 of 100 on Ora's agent-readiness scan
In short
Goldman Sachs scores 26 of 100 on Ora's agent-readiness scan, grade F, read on 9 November 2026. 9 percent of the 132,085 domains Ora ranks score under 20, and the average score is 44. Six agents blocked in front of a page the reading checks mostly pass.
The score
Goldman Sachs is an investment bank based in New York.
- 26of 100, grade FOra's agent-readiness scan of goldmansachs.com, read 9 November 2026.
- 3Discovery, of 12Whether assistants find the company: search presence, Wikipedia and Wikidata, catalogues.
- 17Access, of 50Whether they can read it: the door, text without JavaScript, llms.txt, structured data.
- 9Usability, of 44Whether they can act through it: an MCP server, an API with a specification, sign-in for agents.
- 0Payments, of 0No check in this layer applied to the site.
9 percent of the 132,085 domains Ora ranks score under 20, and the average score is 44.
Ora's one-line summary: Goldman Sachs provides a sitemap, but lacks a published OpenAPI specification.
What an assistant meets at the door
Before an assistant can read, compare or buy, it has to get in. These rows are the door: whether the bot shield lets the main agents through, what robots.txt tells the search, training and user-triggered agents, and whether the home page has text without running JavaScript.
| Check | Result | What Ora saw |
|---|---|---|
| Not blocked by bot detection | Fail, 0 of 2 | Agents blocked: GPTBot, ClaudeBot, ChatGPT-User, PerplexityBot, Google-Extended, Applebot-Extended |
| Agent crawler reachability | Partial, 0 of 2 | Some AI crawlers are blocked - ChatGPT-User: blocked, ClaudeBot: blocked, Google-Extended: blocked, ora-agent: reachable, DeepSeekBot: blocked |
| robots.txt AI crawler policy | Partial, 0 of 2 | AI crawlers mentioned (gptbot) but no explicit Allow directives |
| robots.txt agent-user policy | Pass, 2 of 2 | User-triggered agents explicitly allowed (chatgpt-user) |
| Content without JavaScript | Partial, 1 of 3 | Only 4894 chars of text content, no H1 tag - agents see limited content |
What it can read, and what it can do
The reading layer is what an assistant takes away from the pages: a text file written for it, markdown on request, structured data about the company, a sitemap. The act layer is whether it can do anything through a published interface rather than by clicking through the site.
| Check | Result | What Ora saw |
|---|---|---|
| llms.txt exists | Fail, 0 of 1 | No llms.txt found at /llms.txt or /.well-known/llms.txt. |
| llms.txt formatting | Fail, 0 of 2 | No llms.txt found to evaluate quality |
| Markdown agent docs | Fail, 0 of 1 | Missing markdown content negotiation. Homepage returned text/html when probed with Accept: text/markdown (expected text/markdown), and none of the well-known root paths (/llms.md, /auth.md, /agents.md, /developers.md). |
| Bot-UA markdown serving | Fail, 0 of 1 | No probed bot User-Agent (GPTBot, ClaudeBot, ChatGPT-User, PerplexityBot, Google-Extended, Applebot-Extended, ora-agent, DeepSeekBot) receives markdown when requesting HTML. |
| JSON-LD structured data | Pass, 3 of 4 | JSON-LD Organization with name + description - add url and sameAs/logo/address for full score |
| Organization schema completeness | Pass, 2 of 2 | Organization schema complete with contactPoint and address |
| Sitemap exists | Pass, 2 of 2 | Valid sitemap found at https://www.goldmansachs.com/sitemap.xml with multiple sitemaps entries |
| Metadata completeness | Pass, 2 of 2 | All metadata signals present: canonical URL, lang="en-US", og:image, og:type |
| Page token budget | Pass, 1 of 1 | All 6 measured pages fit an agent context budget (largest ~1K tokens) |
| Check | Result | What Ora saw |
|---|---|---|
| MCP server / manifest | Fail, 0 of 6 | No MCP server or manifest found |
| MCP well-known discovery | Fail, 0 of 2 | No MCP server URL discovered |
| Public API with reachable endpoints | Fail, 0 of 7 | No publicly reachable API surface detected (REST and GraphQL both absent or auth-gated) |
| Public API/docs linked from homepage | Fail, 0 of 3 | No public API or documentation page linked from homepage |
| OpenAPI spec published | Fail, 0 of 7 | No OpenAPI/Swagger specification found |
| Developer portal | Did not apply, 0 of 6 | No developer portal found |
| OAuth 2.0 support | Fail, 0 of 5 | No OAuth 2.0 or OpenID Connect support detected |
| Scoped permissions | Fail, 0 of 5 | No declared OAuth scopes, security schemes, or scoped-permission documentation found |
| Multi-language SDK packages | Did not apply, 0 of 3 | No SDK packages found across language ecosystems |
| CLI tool available | Did not apply, 0 of 3 | No CLI tool found |
| WebMCP support | Fail, 0 of 2 | No WebMCP support detected - no tool-attribute forms and no document.modelContext / navigator.modelContext usage found (scanned 8 of 8 same-origin script bundle(s) for modelContext registrations) |
| Check | Result | What Ora saw |
|---|---|---|
| Wikipedia / Wikidata entity presence | error, 0 of 4 | Wikipedia and Wikidata presence could not be verified - try rescanning |
| Brand name discoverability | Partial, 1 of 3 | goldmansachs.com appears once in brand-name search results for "Goldman Sachs" (position #4 out of 6) |
| Category share of voice | Did not apply, 0 of 6 | Category share of voice is in beta. It measures whether agents surface you for capability and use-case searches (not just your name). |
| Developer resource discoverability | Partial, 2 of 6 | Agent found 2 pages by name but no recognizable developer-resource type. Not found via search: API docs, OpenAPI spec, developer portal, auth docs, MCP server |
The six payment checks did not apply on this read: Ora did not read the site as a shop.
Three fixes to hand to the web team
Ora estimates how many points each open check would return. These are the largest on this read, in Ora's words where it gave them, so a leader can hand the list over without translating it.
- Developer resource discoverability, worth about 7.4 points by Ora's estimate: Agent found 2 pages by name but no recognizable developer-resource type. Not found via search: API docs, OpenAPI spec, developer portal, auth docs, MCP server.
- Public API with reachable endpoints, worth about 7.1 points by Ora's estimate: No publicly reachable API surface detected (REST and GraphQL both absent or auth-gated).
- MCP server / manifest, worth about 6.1 points by Ora's estimate: No MCP server or manifest found.
Our read
Goldman Sachs blocks the agents and publishes a well-made page none of them will see. GPTBot, ClaudeBot, ChatGPT-User, PerplexityBot, Google-Extended and Applebot-Extended were blocked at the door, while robots.txt names gptbot without an explicit allow and permits chatgpt-user, a mixed message. Behind the block the reading layer is in order: the home page gives 4,894 characters, though without an H1, the Organization block is complete, the sitemap index is valid and the metadata carries all four signals, so the Access layer still reads 17 of 50; there is no llms.txt. The act layer is empty, with no public API, no developer portal and neither an MCP server nor OAuth discovery, so the Usability layer reads 9 of 44. The Discovery layer is 3 of 12: the domain appears once, fourth of six, in a search for its own name, and the Wikipedia check could not be verified. An investment bank that blocks the assistants its clients now use has made a choice; the F records it.
Test your own site
Run the same scan on your own domain at ora.ai. The score, the layers and the fixes on this page come from its public result for goldmansachs.com; a scan of a public site takes a few minutes and needs no account.
- Scan your domain on OraThe same checks, the same scoring, on your own site.
- The self-check, 18 questionsWhat to ask for, in business terms, before the scan.
- All the 360 checksThe other companies in this series, newest first.
Sources
- Ora report for goldmansachs.com (ora.ai/score/goldmansachs.com), read 9 Nov 2026. Vendor claim.
Last reviewed