Researched account
Northwind Clinical Systems
northwindclinical.com · Boston, MA · health-tech platform · B2B SaaS
Fit score
0
of 100
Targeting
Representative field. The questions, the survivor share and the reason grammar are the product’s; the companies behind the bars are illustrations, not client data.
Each stage, in detail
Stage 01
Discovery runs live search against queries built from your ICP and the segment you define. The sources are named rather than implied: organic web search, and Clutch profiles where the market is service providers. There is no purchased list and no resold database anywhere in the pipeline.
Every candidate is resolved to a real company, normalized, and deduplicated against what we already know. The raw discovery trail is kept, so a later decision can be inspected rather than guessed at.
No research happens here. The job is to establish that a real company exists before any expensive work begins — and a search result is never allowed to write itself into a company’s canonical homepage field, because a directory page is not a company.
Stage 02
Cheap deterministic scoring runs before any model does. Industry and segment fit, the keywords and exclusions you set, structural junk checks, and geography where you have chosen to constrain it decide the obvious cases.
Clear fits are shortlisted. Clear non-fits are discarded for that client. Only the genuinely ambiguous move to triage. In development running, 91% of candidates were discarded here, before a cent of research money was spent on them.
Every score is client-specific and carries its reasons, so you can see why an account moved forward — and why one stopped, which is the half most tools do not keep.
Stage 03
Only ambiguous accounts escalate to AI review. The model resolves uncertainty from evidence already gathered, with bounded prompts and stored reasoning.
Triage is not enrichment and it is not browsing. It is a classification step that protects the expensive parts of the pipeline.
The output is still reviewable: reasons, confidence, and the evidence that made the account worth or not worth continued spend.
Stage 04
Only shortlisted accounts get researched, because this is the step that costs real money — around four fifths of the variable cost of running LeadAtomic sits in this one call.
The research gathers what the company does and who they sell to, headcount and hiring, funding signals, sales-stack signals, recent news, risk flags, and contact addresses with a verification verdict on each. Weak or missing evidence stays visible instead of being smoothed over.
Failures become explicit states, not silent drops, so the queue remains honest and debuggable.
Stage 05
The research becomes an outreach draft that names the signal it was written from, rather than filling a template.
Approval is on when you start and stays on until you turn it off. What you cannot turn off is the floor: a message that cannot carry a sender identity and a working unsubscribe link does not send at all.
Sending runs behind mailbox warm-up, enforced volume ceilings, and bounce brakes. Replies come back through your connected mailbox and stay tied to the account.
What lands in your queue
Researched account
Northwind Clinical Systems
northwindclinical.com · Boston, MA · health-tech platform · B2B SaaS
Fit score
0
of 100
Ruled out
Drafted opener · awaiting your approval
“Congratulations on the June round. Three of your nine open roles are in platform security, which usually means the audit conversation is already happening internally rather than being something someone still has to raise.”
written from: funding_signal, hiring_now
What you control
01
Market and segment, ICP, exclusions, keywords, and brand voice
02
Manual, semi-automatic, or rules-based outbound control
03
Score thresholds, evidence expectations, and approval posture
04
Which mailbox sends, and therefore which domain carries the reputation
We launch on 1 October 2026. Join the waitlist for first-come-first-served early access, with launch pricing locked for your first 12 months.