Detect a company's tech stack from its website
Builds an evidence-backed technographic profile per account: up to eight tools with category, the exact place on the site each was spotted, and a uses-versus-integrates flag. Cheaper than technographic data vendors for targeted lists, and the evidence field makes every entry auditable and quotable in copy.
You are a technographics analyst running inside an enrichment pipeline. You identify the tools a company visibly uses by reading evidence on their website — careers pages naming a stack, integration and partner pages, case studies mentioning platforms, footer badges, docs, and job descriptions. You report only what the site itself shows; you never rely on general knowledge about what companies 'typically' use.
Work from the company website you are given (a URL to browse or scraped page content). Check, in order of reliability: job postings, integration/partner pages, documentation, case studies, and page footers.
Rules:
- Every tool you report must have a stated location on the site where you saw it.
- Distinguish tools they USE from tools they INTEGRATE WITH or sell to users of — a Salesforce integration page does not mean they run Salesforce internally.
- Do not report inferred tools ('probably uses AWS'). No evidence, no entry.
- Category labels: crm, marketing-automation, analytics, data-warehouse, sales-engagement, support, payments, cloud, framework, other.
OUTPUT CONTRACT — follow exactly:
Return a single JSON object:
{"tools": [{"name": "<tool name>", "category": "<one label from the list>", "evidence": "<under 12 words: where on the site you saw it>", "relationship": "<uses | integrates>"}], "confidence": "<high | medium | low>"}
- Return ONLY the JSON. No markdown fences, no commentary.
- Maximum 8 tools, most sales-relevant first.
- If the site is unreachable or shows no identifiable tools, return exactly: {"tools": [], "confidence": "low"}
- Never list a tool without evidence you can point to.How to use it
- 1
Copy the prompt into Claude, ChatGPT, or any LLM — or into a Clay Claygent column.
- 2
In Clay, map your domain or website column as the input; give the Claygent permission to browse multiple pages per row.
- 3
Validate a 20-row sample against a source like BuiltWith or job postings — the evidence field should let you confirm every entry in seconds.
- 4
Filter on the tools you displace or integrate with, and merge the evidence into your opener ('your SDR job post mentions Outreach').
Best practices
Job postings are the highest-signal source — companies name their real stack when hiring for it. Weight those entries over footer badges.
Never treat an integrates entry as displacement opportunity; selling against a tool they merely connect to reads as sloppy research.
Combine with a Clay technographics provider in a waterfall: run this prompt only on rows the provider returned empty, and save the credits.
Example: what this looks like in practice
A team selling a Salesforce-native forecasting tool needs to know which of 5,000 mid-market accounts actually run Salesforce. Their data vendor covers 60%. They run this prompt on the 2,000 unknowns: 740 come back with Salesforce under uses, most citing job postings ('Salesforce admin experience required' on the careers page). The evidence string goes straight into the email: 'Your RevOps posting asks for Salesforce flow experience, which usually means forecasting still lives in spreadsheets next to it.' The integrates flag saves them from 300 embarrassing sends to companies that merely list a Salesforce app-exchange integration.
Best fit
This prompt is one gear in a bigger machine. We orchestrate 20+ tools into outbound systems our clients own — and guarantee the results.
Apply for a Pilot Spot → →Frequently asked questions
Yes, for a meaningful share of accounts. Companies leak their stack constantly — job posts naming required tools, integration pages, docs, case studies, footer badges. An AI agent reading those sources with an evidence requirement recovers reliable technographics on 30–50% of rows vendors miss, at a fraction of the cost. It complements vendors; it doesn't fully replace them.
More personalization at scale (clay & claygents) prompts
Generate a cold email first line from scraped research data
The workhorse of AI-personalized outbound: converts whatever research your enrichment waterfall produced into one 12–22-word observation-led first line per row. The swap-test rule and fact-invention ban are what separate lines that read researched from the AI slop prospects now delete on sight.
Turn a Claygent website scrape into a personalized one-liner
Produces a sub-25-word, website-specific opening line for every row in a Clay table — no human in the loop. The output contract forces plain text only and a NOT_FOUND fallback, so unreachable sites and thin homepages return a filterable sentinel instead of a hallucinated compliment that ends up in a prospect's inbox.
Analyze a pricing page for sales-relevant intelligence
Extracts a structured pricing profile from any pricing page: model, tier count, exact displayed prices, free-offer type, and a one-line read on their sales motion. At scale it segments prospect lists by how companies charge — which predicts budget, buying process, and which of your angles will land.