Skip to content
Index / Methodology

Methodology · 2026-11-17

How the index works

Every number on this site comes from a public page. If we couldn't find it, the field is blank.

1. Discovery

The index covers three pillars of AI engineering work: AI automation, AI internal tools, and custom AI apps and agents. An agency is placed in a pillar only when its own website repeatedly describes that work (for custom AI apps: at least five mentions such as RAG, LLM application development, LangChain or AI agent development, plus an explicit LLM/RAG/agent phrase). Agencies can sit in several pillars. Candidates come from public directories (aiautomationdb.com, aiagencymap.com, theexpertroster.com, partnerhire.com), the official n8n and Zapier partner directories, and public "best agencies" articles. Domains are de-duplicated.

2. Verification

A crawler fetches each homepage plus contact, about, pricing and case-study pages. A listing passes if the site is live, automation/AI is a core service, the site shows 2025–2026 activity, and it reads as an agency rather than a single SaaS product. Listings are re-checked weekly; dead sites are flagged and removed.

3. Enrichment

We extract service headings, case-study titles, sentences that state outcomes, logo alt-text in client sections, published prices, team size and founders. Outcomes are shown as quotes attributed to the agency; we do not verify them.

4. The Cofounderbase Score (0–100)

ComponentMaxRule
Stack depth25Core tools of each pillar named on the agency site: automation (n8n, Make, Zapier, Power Automate…), internal tools (Retool, Appsmith, Budibase, ToolJet, Superblocks) and AI app frameworks (LangChain, LlamaIndex, LangGraph): 8 pts each. Supporting tools (OpenAI, Claude, Pinecone, CrewAI, voice-agent and no-code builders): 4 pts each. +3 for 4+ tools.
Evidence25Case-study pages found (4 pts each, max 3), named case studies (3 each, max 3), stated outcomes (2 each, max 2), linked review profiles (+4).
Certification2020 if found in an official partner directory (n8n, Zapier); 8 if a partnership is only self-reported on the agency site.
Transparency15Published pricing (6), team size (3), named founder (3), public contact email on own domain (3).
Freshness15Footer/copyright year 2026 = 15, 2025 = 10; otherwise 6 (passed our 2025–26 activity check).

The score measures public evidence and transparency — not the quality of an agency's work. Paid plans never change the score.

Current distribution

0–1959
20–39412
40–59426
60–79169
80–1009

5. Labels

verified found in an official partner directory. self-reported claimed on the agency's site only. Paid listings get a "claimed" label; paying never changes the score.

Known limitations

Get matched with 3 vetted agencies

Describe the project. We read every brief and introduce you to three agencies from this index that fit it, usually within 2 working days. Free for buyers; agencies can't pay to be picked.

Add details (optional): company, type, budget, timeline
CompareOpen shortlist