Product
Everything between “who should we talk to?” and a worked pipeline.
Five connected capabilities, one data model, and a provenance trail on every field.
1. Business discovery
Discovery jobs over reviewed sources
Define the location, category and limit. The job is queued, processed by background workers with per-source rate limits, and reports its outcome per source: pages fetched, records created or updated, duplicates found, errors and retries.
Cost preview
A credit estimate before the job runs. Credits are reserved on queue and settled against actual output.
Cancellation and retry
Cancel a running job safely or retry only the items that failed.
Honest coverage
Each source shows its coverage notes and retention rules. Nothing claims to be complete.
2. Contact enrichment and verification
Normalize, verify, and keep the original
Phones are parsed with country awareness. Emails are syntax and mail-server checked, with disposable domains flagged. Domains and social URLs are canonicalized. The raw observed value is kept beside the normalized one.
Six field states
Verified, unverified, stale, expired, missing, conflicting. Conflicts are shown, never silently resolved.
WhatsApp with evidence
Marked confirmed only from an explicit link or label on the business's own site, or a team member's assertion after contact.
Freshness windows
Fields age into stale after your configured window and can be refreshed from their source.
3. Website analysis
Observed facts about a public website
A sandboxed fetcher reads a handful of public pages, respects robots directives, never executes scripts, and records detections with evidence, confidence and the analysis date.
Detections
No reachable website, no HTTPS, no contact form found, no booking flow found, no online ordering found, no WhatsApp link found, missing viewport meta.
Opportunities
Suggestions derived from detections, each linked to the evidence that produced it.
Careful wording
The product says 'no X was detected', because a missed detection is not proof that X does not exist.
4. Lead scoring and deduplication
Explainable fit, cautious merging
A fit score from weights you control, shown as a breakdown and separated from data confidence. Duplicates are found by exact domain, phone and email first, then name-and-location similarity; ambiguous matches go to a review queue and merges can be undone.
Your criteria
Weight contact completeness, verification, website opportunity, industry match, location match, freshness and source reliability.
No name-only merges
Similar names never merge automatically. You decide, with both records side by side.
Merge history
Every merge keeps a snapshot so an incorrect merge can be reversed.
5. Lead management
Lists, pipeline, activities and exports
Save businesses to lists, move leads through configurable stages, assign owners, log calls, WhatsApp messages, emails and meetings, and export authorized records within your plan's limits.