Describe what you sell
Share your website or describe the problem you solve, who benefits and what makes a company a useful fit.
For examplecrawlkit.sh
Web data infrastructure for AI agents and research tools.
Find companies that actually need what you sell.
Tell CrawlKit what you sell. It researches the live web to find companies that could actually be a fit.
From your product to a reason to talk.
Customer discovery starts with a need, not just an industry filter.
Research the public web
Look beyond a company description. Product pages, documentation and hiring requirements help connect an account to a concrete use for your product.
Review the fit and its evidence
Separate what a source says from what it suggests. Keep the source link and the reason for following up, including evidence that weakens the match.
What makes a useful signal?
- Product pages
- A capability that relies on a problem your product solves.
- Documentation
- A technical dependency, integration or workflow requirement.
- Careers pages
- A team investing in relevant skills. A hiring signal is context, not proof of a purchase.
- Changelogs and announcements
- A dated change that makes an account worth a fresh look.
A real source. A careful conclusion.
A worked research example using public Dify sources, reviewed on September 27, 2026. Not a customer case study or an automated CrawlKit result.
Research question
Which AI platforms have a visible use for structured web content?
Observed evidence
Dify’s Knowledge Pipeline page describes bringing web content into knowledge bases for AI applications.
Source: Dify Knowledge Pipeline ↗Dify’s June 17, 2024 announcement documents a Firecrawl integration for importing website content.
Source: Dify integration announcement ↗Difydify.ai · AI application platform
What this suggests
Web data is relevant to the product. That is a reason to research the account, not proof that it needs CrawlKit.
What weakens the fit
The published integration shows an existing solution. An unmet need or reason to switch has not been established.
What to verify next
Whether there is a current gap in coverage, output formats or data freshness that another provider could address.
No buying intent, relationship with CrawlKit or conversion outcome is asserted.
Apply the same approach to your market.
Practical research scenarios, not customer success claims.
Developer tools
Find products whose documentation exposes a relevant workflow or dependency. Check whether your tool solves a missing part rather than duplicating an existing integration.
B2B software
Connect a launch or hiring requirement to the problem your software solves. Validate the need before treating the company as an opportunity.
Specialist services
Research teams taking on work related to your expertise. A new role may indicate investment, but could also mean they plan to handle the work internally.
Customer discovery, explained.
Start with your website or a description of what you sell. The workflow connects your product to relevant companies through public web research, then explains the potential fit with supporting sources.
No. Customer discovery focuses on public web research, not only filtering a static company list. Product pages, documentation, company websites and hiring information can provide context for why a company may be relevant.
It means there is a supported reason to investigate a company further. A matching technology or product feature is not proof of buying intent, budget or an unmet need. Review the evidence and confirm these details before reaching out.
No. The Dify example on this page is a manually reviewed walkthrough based on linked public sources. It is not an automated CrawlKit result, an endorsement or evidence that Dify uses CrawlKit. No customer outcome or match score is claimed.
B2B founders, sales teams and service providers can use public signals to narrow a research list. The best fit is a product whose use can be connected to visible company activities, such as a software integration, a hiring requirement or a product launch.
Yes. Search, Scrape, Extract, Crawl, Watch and MCP remain available. You can build your own research workflow with the same CrawlKit platform.
One platform. Two ways to use it.
Research potential customers, or build your own workflow with the live web.