3.2x Pipeline in 90 Days with an AI Prospecting Engine Built on Seamless AI + Clay
A Series B SaaS company's SDRs were losing 20 hours a week to manual list building, with contact data that was only 61% accurate. SiyaSphere built a waterfall enrichment engine on Seamless AI and Clay that scores ICP fit automatically and routes accounts into Outreach sequences by tier. Within 90 days, pipeline was 3.2x the previous quarter and cost per qualified lead fell 44%.
- Pipeline vs. previous quarter
- 3.2x
- Pipeline vs. previous quarter
- Contact data accuracy (from 61%)
- 94%
- Contact data accuracy (from 61%)
- Cost per qualified lead
- -44%
- Cost per qualified lead
The challenge
- SDRs spent 4+ hours a day building prospect lists by hand from LinkedIn.
- Contact data went stale quickly, holding contact rates at 8%.
- There was no systematic way to prioritise accounts by ICP fit, so pipeline was inconsistent and reactive.
What we built
- 1
Real-time contact discovery
Deployed Seamless AI to find verified B2B emails and mobile numbers in real time, replacing manual LinkedIn research.
- 2
Waterfall enrichment in Clay
Built Clay workflows that auto-enrich every new account from LinkedIn, Clearbit, and BuiltWith, with Apollo.io as a fallback whenever Seamless AI confidence dropped below 80%.
- 3
AI-driven ICP scoring
Added an AI scoring formula in Clay that grades each account's ICP fit and tags it into a tier.
- 4
Automated CRM sync and sequencing
Pushed enriched, tagged records into Salesforce and routed them to the right Outreach sequence based on ICP tier, with no manual hand-off.
Results
- Manual list building cut from 20 hours/week to zero
- Contact data accuracy rose to 94% (up from 61%)
- SDR outreach volume up 180%
- Pipeline generated in 90 days reached 3.2x the previous quarter
- Cost per qualified lead reduced by 44%
Key takeaway
Single-source enrichment leaves gaps. A waterfall (primary provider, then fallbacks gated on confidence scores) is what makes AI prospecting data accurate enough to automate sequencing end to end.