# 3603: Agent-Direct Score F (23.6%)

Site: https://www.3603.co.za/
Report: https://directindex.org/r/3603.co.za
Scanned: 2026-11-14
3603 is very hard for agents to use.
An AI assistant can open 3603's website. It can't find a menu or price list, so it answers from other sites. It can't find a link to book or order online.

## Direct links
- Phone: +27 688706745 (from Overture Maps)
- Address: 1 Clifton Rd, 3610, Hillcrest (from Overture Maps)
- Is this right? https://directindex.org/about?dispute=3603.co.za#contact

## What assistants answer
Asked on 2026-11-15 with web search on. 67% of the answers pointed to the business's own site or to a vendor working for it. That share is the Cited part of the score.
| Assistant | Official website? | Where to book? | Where to order? |
|---|---|---|---|
| ChatGPT | own site | other site: dinestar.co.za | own site |
| Gemini | own site | other site: www.findmy.co.za | own site |

Named in none of the 44 answers to 22 everyday questions about Durban.

Named instead. In the answers that didn't name 3603, the places named most often, and in how many answers.
- Dukkah Restaurant & Bar (https://directindex.org/r/dukkah.co.za): in 9 answers
- The Chefs Table (https://directindex.org/r/thechefstable.co.za): in 9 answers
- The Grill Room at The Oyster Box (https://directindex.org/r/oysterboxhotel.com): in 9 answers
- Spiga (https://directindex.org/r/spiga.co.za): in 6 answers
- Afro’s Chicken Shop (https://directindex.org/r/afroschicken.co.za): in 3 answers
See who was named: https://directindex.org/city/durban/recommended

**What ChatGPT said, Official website? (2026-11-15):** https://www.3603.co.za/ (thesaunter.co.za (https://www.thesaunter.co.za/listing/3603/?utmsource=openai))

**What ChatGPT said, Where to book? (2026-11-15):** Book a table online via 3603’s Dinestar reservation page (https://dinestar.co.za/en/restaurants/1246-3603-durban). (dinestar.co.za (https://dinestar.co.za/en/restaurants/1246-3603-durban?utmsource=openai))

**What ChatGPT said, Where to order? (2026-11-15):** Use 3603’s official website to check ordering options: https://3603.co.za/ (https://3603.co.za/). (3603.co.za (https://3603.co.za/))

**What Gemini said, Official website? (2026-11-15):** The official website of 3603, a restaurant in Durban, South Africa, is www.3603.co.za.

**What Gemini said, Where to book? (2026-11-15):** You can book a table online at 3603 Cafe & Bistro in Gillitts, Durban, through the FindMy platform. Here's the link you should use: https://www.findmy.co.za/food/restaurant/cafe/gillitts/1/restaurant-3603-gillitts

**What Gemini said, Where to order? (2026-11-15):** You can order food online from 3603 Cafe & Bistro in Durban, South Africa, by visiting their official website. Here is the link you should use: https://www.3603.co.za/

## Why the assistants send people elsewhere
Why (2026-11-15): ChatGPT directed users to Dinestar and Gemini sent them to FindMy because the business website lacked any booking link and displayed almost no readable text without JavaScript.

## What to fix first
- **No structured data for the business.** [no_structured] (+15 points) Add schema.org JSON-LD for the business (Restaurant, LocalBusiness or the closest type, with name, address, telephone and openingHoursSpecification, plus servesCuisine and acceptsReservations). Every assistant reads this block first, and the fix kit below has it filled in. Passes when the home page has schema.org JSON-LD that names the business and its type. Guide: https://directindex.org/guides/structured-data-restaurant
- **No booking or ordering link on the website.** [no_channel] (+15 points) There's no link on the site for an agent to follow. If you take bookings or orders through a platform, link to it from the site so agents can find it. Add a booking or ordering page on your own domain. Passes when the home page links a booking or ordering page on the site's own domain, or booking or ordering software. Guide: https://directindex.org/guides/book-direct-on-your-own-website
- **No menu or price list found.** [menu_none] (+7 points) Agents answer 'what do they offer?' from marketplaces instead. Add a page with the menu or price list and link to it from the homepage. Passes when the home page links a menu or price list as an HTML page on the site. Guide: https://directindex.org/guides/menu-pdf-ai
- **Almost no text without JavaScript.** [no_text_nojs] (+2 points) Crawlers that don't run scripts see an empty page. Put the menu, hours and address in plain HTML. Passes when the home page carries at least 500 characters of text without JavaScript. Guide: https://directindex.org/guides/website-blank-without-javascript
- **No llms.txt.** [no_llms_txt] (no points, but worth doing) Optional for now and cheap to add. It's a short text file that describes the business for language models. Passes when /llms.txt answers with a short text file (not scored). Guide: https://directindex.org/guides/llms-txt-for-a-business

## Fix kit: What to paste

Structured data (JSON-LD). Paste it as one block in the page's head, or in the builder's custom code box. Change anything that's wrong.

```html
<script type="application/ld+json">
{
  "@context": "https://schema.org",
  "@type": "Restaurant",
  "name": "3603",
  "url": "https://www.3603.co.za/",
  "telephone": "+27688706745",
  "address": "1 Clifton Rd, 3610, Hillcrest",
  "geo": {
    "@type": "GeoCoordinates",
    "latitude": -29.79605270537658,
    "longitude": 30.790260218098588
  }
}
</script>
```

llms.txt. Save as /llms.txt at the root of the site.

```
# 3603

> restaurant in 1 Clifton Rd, 3610, Hillcrest, Durban. Phone +27 688706745.

## About

- Website: https://www.3603.co.za/
- This file follows llms.txt (https://llmstxt.org): a short plain-text description for language models.
```

## Already right
- **robots.txt lets the assistants in.** There's no robots.txt, which means everything is allowed.
- **AI crawlers get the page.** GPTBot, ClaudeBot and PerplexityBot all got a normal response.

## Where it stands
224 of 292 in Durban. Ahead of 23% of the other sites scanned in this index.

| What the top 10% of the index does | Top 10% | This site |
|---|---|---|
| Take bookings or orders on their own domain | 93% | no |
| Publish the menu or price list as HTML | 97% | no |
| Let the assistants in (robots.txt) | 100% | yes |
| Structured data with address, phone and hours | 27% | no |
| Have an llms.txt | 37% | no |


See what an AI agent reads on this site: https://directindex.org/r/3603.co.za/agent-view

## Businesses like 3603 in Durban

| Group | Sites | Avg score | A or B | This site |
|---|---|---|---|---|
| Businesses of the same kind in Durban | 292 | 47% | 17% | 23.6% |

With the three fixes above (No structured data for the business; No booking or ordering link on the website; No menu or price list found), the score would be about 91%. Against today's scores in Durban, that would rank about 9th of 292 (it is 224th today). The others' scores move too, so this is an estimate.

## The direct record
What agents get for 3603, as JSON: https://directindex.org/api/direct/3603.co.za
- Book: none on record
- Order: none on record
- Menu: none on record
- Hours: none on record
- Phone: +27688706745

## Can an agent transact?
Measured, not scored (not part of the Agent-Direct Score, spec 1.1).
- UCP profile (/.well-known/ucp, Google and Shopify's Universal Commerce Protocol): no
- ACP discovery document (/.well-known/acp.json, OpenAI and Stripe's Agentic Commerce Protocol): no
- MCP server card (/.well-known/mcp/server-card.json): no
- Ordering vendor that takes agent orders (Square, Toast, or a vendor whose terms allow agents): no

## Score breakdown
| Component | Points | Available |
|---|---|---|
| Reachable | 13 | 15 |
| Readable | 0 | 25 |
| Actionable | 0 | 15 |

Scored on what the scan measures for every site, the same way for everyone. The action test and what the assistants answer are shown beside the score, never in it (specification 1.1, 18 November 2026).

### Who can read the site, by kind of agent
| Kind of agent | What the scan saw | In the score |
|---|---|---|
| Search crawlers (OAI-SearchBot, Claude-SearchBot, PerplexityBot) | let in by robots.txt | 10 points, the two rows together |
| Fetchers that read a page for a person (ChatGPT-User, Claude-User, Perplexity-User) | let in by robots.txt | 10 points, the two rows together |
| Training crawlers | let in by robots.txt | not scored |
| Requests under crawler names from Cloudflare's network | got the page | not scored |
| The scanner's own signed request | got the page | 2 points: the text without JavaScript |

## Agent readiness (Cloudflare)
Level 0 of 5, Not Ready. Cloudflare's report: https://isitagentready.com/3603.co.za
- Link headers on the homepage: no (how: https://isitagentready.com/.well-known/agent-skills/link-headers/SKILL.md)
- sitemap.xml: no (how: https://isitagentready.com/.well-known/agent-skills/sitemap/SKILL.md)
- AI bot rules in robots.txt: no (how: https://isitagentready.com/.well-known/agent-skills/ai-rules/SKILL.md)
- Content Signals in robots.txt: no (how: https://isitagentready.com/.well-known/agent-skills/content-signals/SKILL.md)
- Markdown on Accept: text/markdown: no (how: https://isitagentready.com/.well-known/agent-skills/markdown-negotiation/SKILL.md)
- llms.txt: no (how: https://isitagentready.com/.well-known/agent-skills/llms-txt/SKILL.md)
- API catalog (RFC 9727): no (how: https://isitagentready.com/.well-known/agent-skills/api-catalog/SKILL.md)
- MCP server card: no (how: https://isitagentready.com/.well-known/agent-skills/mcp-server-card/SKILL.md)
- A2A agent card: no (how: https://isitagentready.com/.well-known/agent-skills/a2a-agent-card/SKILL.md)
- Agent Skills index: no (how: https://isitagentready.com/.well-known/agent-skills/agent-skills/SKILL.md)
- OAuth or OIDC discovery: no (how: https://isitagentready.com/.well-known/agent-skills/oauth-discovery/SKILL.md)
- OAuth protected resource metadata: no (how: https://isitagentready.com/.well-known/agent-skills/oauth-protected-resource/SKILL.md)
- auth.md: no (how: https://isitagentready.com/.well-known/agent-skills/auth-md/SKILL.md)
- Web Bot Auth key directory: no (how: https://isitagentready.com/.well-known/agent-skills/web-bot-auth/SKILL.md)
- ARD capability manifest: no (how: https://isitagentready.com/.well-known/agent-skills/ard/SKILL.md)
- DNS-AID records: no (how: https://isitagentready.com/.well-known/agent-skills/dns-aid/SKILL.md)
- WebMCP tools on the page: no (how: https://isitagentready.com/.well-known/agent-skills/webmcp/SKILL.md)
- JSON-LD for the business: no

Own this business? Claim it, correct what agents are told, recheck a fix, or hand the list to your assistant: https://directindex.org/r/3603.co.za/owner
Methodology: https://directindex.org/methodology · Rescan now: https://directindex.org/scan?url=3603.co.za&force=1
