From a company name to a tracked conversation
Eight stages run in order. Stages 1–4 are automated intelligence; 5–8 are the seller's execution loop.
Discover
DiscoverCompaniesDialogTarget a company by name or website, or use Discover Companies to find prospects.
Research
analyzeCompanyAI runs web research: business model, tech stack, pain points, recent news, and decision makers.
Score
IcpFitCard · OpportunityScoreCardICP Fit (direct-customer) + Opportunity Score (readiness, likelihood, deal size, sales-motion viability, research economics).
Recommend
classifySalesMotionA deterministic classifier turns the scores into one action — the Recommended Sales Motion.
Prioritize
Home · PipelineDashboard cockpit, pipeline, and lead scoring surface what to act on today.
Personalize
Outreach · MeetingPrep · Solution DesignerAI generates outreach (email + LinkedIn sequence), meeting prep, tailored solution, and proposal.
Engage
sendEmail · sendOutlookEmail · LinkedInSequenceSend through your own Gmail/Outlook; LinkedIn uses an open-and-copy pattern (ToS-safe).
Track
checkForReply · workflowsReply detection, opportunity stages, activity timeline, and automated follow-up workflows.
How scores become a single action
The classifier is deterministic — no extra AI cost, fully reproducible — and respects the ICP engine's own verdicts.
Deterministic rules + guardrails
- Fit + timing + relationship + strategic value
- Opportunity Type caps & partnership preference
- Evidence Confidence downgrades weak calls
- Opportunity Score only breaks ties at the margins
What the seller actually sees first
One verb, not a dashboard of numbers. The scores sit underneath as supporting detail.
What a rep reads, top to bottom
Action first, evidence last — so an ordinary salesperson acts instead of interpreting scores.
Design principle: SalesLensAI decides what should happen; the seller's existing channels (Gmail, Outlook, LinkedIn, CRM) handle delivery. The next strategic step is closed-loop tracking — what was recommended → what action was taken → what happened — so the recommendation engine learns from outcomes.