A precise description of the kind of company a business sells to best — typically expressed as industry, size, geography, technographics, and exclusion criteria. In LeadAtomic, the ICP drives deterministic scoring before any AI call. A vague ICP produces noisy results regardless of tooling; sharpening the ICP is usually the highest-leverage improvement available.
The process of attaching context to a discovered company: contactability, company summary, signals (hiring, funding, news), technographics, and other evidence relevant to outreach. LeadAtomic enriches only after a lead has passed qualification, so the enrichment budget is reserved for leads worth the spend. Weak or missing evidence stays visible rather than being smoothed over.
Assigning a fit score to a lead so the pipeline can decide whether to shortlist, discard, or escalate. Scoring in LeadAtomic is client-specific (the same company can score differently for different clients) and carries reasons so a careful operator can audit the decision. Cheap deterministic checks run first; AI triage resolves only the ambiguous middle.
An AI system whose decisions can be inspected by the person relying on them. In a lead-intelligence context, explainability means every shortlist, discard, or escalation carries the inputs and reasoning behind it. A score without an explanation is just a number you have to trust. LeadAtomic treats explainability as a precondition for shipping AI into the pipeline.
Rule-based qualification that runs before any AI call. Geography, industry fit, exclusions, keywords, and structural junk checks handle the obvious cases cheaply. Deterministic-first is a discipline: AI is then escalation for ambiguity, not the default engine. Reversing the order is expensive, opaque, and tends to amplify noise.
A long-lived signal attached to a sending domain and mailbox identity, used by inbox providers to decide where messages land. Bad sends compound — high bounce rates, spam complaints, and low engagement degrade reputation across all future messages, not just the ones that caused the problem. LeadAtomic treats sender reputation as an asset to protect by default, not a metric to recover after damage.
Whether a sent message actually reaches the recipient's inbox rather than spam or junk. Deliverability is the joint output of sender reputation, authentication (SPF, DKIM, DMARC), content patterns, and recipient engagement. The fastest way to ruin deliverability is to send low-signal messages to people who do not want them; the slowest way to recover it is to do that repeatedly.
The principle that a company has one objective identity (the canonical Company record) and many client-specific interpretations of that identity (ClientCompany records). LeadAtomic strictly separates the two: canonical data is shared and amortised across clients, while client decisions stay isolated. Blurring the boundary is treated as a data-integrity hazard.
Scoring a company specifically on how well it matches a client's defined ICP, as distinct from generic firmographic scoring. ICP fit scoring is client-scoped (it produces different scores for the same company across different clients) and is the first gate LeadAtomic applies before enrichment runs.
Behavioural signals — research activity, content consumption, technology adoption — that suggest a company may be actively considering a purchase. Intent data is a useful but noisy signal. LeadAtomic does not centre intent data; the product is built around verified company candidates and ICP fit, with intent-style signals appearing as supporting evidence when available, not as a primary driver.