What was great about it is, the pipeline is the right shape. It doesn't ask an LLM to "find me leads." It works in stages: Business ā Market ā Buyer persona ā Fit ā Account ā Signal. Each stage narrows the one before, so the shortlist comes from a defined persona and value proposition rather than a vague prompt. For a vertical like ours, that's what makes the output usable.
Citations make it auditable. Every claim in a discovery brief links back to a public source. That's how you deal with hallucination in practice: I can check a figure before I put it in front of a prospect. [e.g. "I spot-checked a few claims about one agency's branch expansion against Companies House and their website, and they held up."]
It sticks to public data. No scraped personal data or grey-market contact lists, which matters for GDPR. Our own customers are letting agents who deal with data compliance every day.
The signal layer adds timing. Weekly alerts on leadership changes, hiring and expansion help with when to reach out, not just who. Agencies changing software usually follow a trigger like that: a new branch, an acquisition, or a new finance hire.
What would make it great for technical teams:
An API or webhooks so briefs and signals can go straight into our CRM or internal tools without copy-paste
Structured export (JSON/CSV) with each claim's source URL and retrieval date kept
Freshness and confidence markers on each claim, so stale data is obvious
UK registry filters such as Companies House SIC codes (68310 estate agencies, 68320 property management), incorporation date and filing history
De-duplication against existing CRM accounts so we don't research customers we already have
Overall, this is grounded, verifiable research rather than generic AI summaries. With an API it would be a strong data source for any B2B team in a vertical market. Nice work, Daniel.
Thanks Cham, really appreciate the feedback. Great to hear you'd already tried the usual stack (Apollo, Sales Nav, Clay, ChatGPT) and that starting from the value prop is what made the difference. That's exactly why we built it that way. You could think about the competitive stack as :
Apollo / ZoomInfo
ā Who can I contact?
Google / LinkedIn / company websites
ā What can I find out?
ClientClarity
ā Who should I care about, and why?
Being able to check a figure before it goes in front of a prospect is the whole point of the citations, so good to know that's what landed for you.
Your list is useful too. CRM de-duplication and a "what's changed since last run" refresh are the ones I'm looking at first. The API and structured export make sense for a technical team like yours.
Would you be up for a quick chat? I'd like to hear how you'd want the Companies House filters to work in practice.