A CSV is not a pipeline. Turn a domain export into an auditable workflow that your sales team can repeat and improve.
Stage 1: normalize the input
Start with source, account name, domain, country and CRM ID. Remove duplicate domains for scanning, but do not destroy separate CRM accounts that legitimately share infrastructure. Validate domain syntax and strip web URLs to the hostname before passing them to the scanner.
account_id,company,domain,source
AC-101,Example Company,example.com,trade-show
AC-102,Example Division,example.org,crm
These rows are fictitious examples. Do not treat them as real prospect leads.
Stage 2: enrich with public infrastructure evidence
Scan the domain list in supported batches. Preserve MX and SPF observations, DMARC policy, recognized email-security provider, workspace inference, confidence and observation date. Keep a distinct null or unknown value rather than replacing it with "none".
For a pre-existing prospect list, use Scan Prospects. If the goal is to discover previously unknown accounts, use Find Leads and its current available dataset filters.
Stage 3: review the highest-risk classifications
Check records where a gateway hides the mailbox platform. Review mismatched Autodiscover and DKIM clues, unknown vendors and failed DNS lookups. A salesperson should always be able to open the raw evidence for an important technical claim.
Stage 4: map segments to offers
| Segment | Candidate offer | Research needed before outreach |
|---|---|---|
M365 indicators + DMARC p=none |
Authentication planning | Confirm company profile and DMARC ownership |
| Recognized gateway MX | Gateway management | Confirm service fit and procurement contact |
| Unknown backend | General mail-architecture discovery | Do not assert Microsoft or Google |
Do not treat an empty gateway field as proof that no protection exists. API-based controls may not be publicly fingerprintable.
Stage 5: enrich decision makers separately
A domain lookup does not reveal a personal email address or the identity of a buyer. Use lawful and appropriately sourced contact research, record where contact data came from and follow applicable outbound-marketing requirements. Keep this contact layer distinct from technographic observations.
Stage 6: send, learn and refresh
Run a small controlled outreach batch. Log the service hypothesis, message variant, reply quality, meetings and rejected assumptions. Rescan before reusing very old evidence. Your CRM should contain both the account and the reason someone thought it was worth contacting, including the limitations of that reason.
Define handoff ownership
The scanning operator owns normalized domains and verified infrastructure observations. A sales-operations analyst owns company deduplication and CRM mapping. The account owner owns fit, contact validation and the outreach decision. This split matters: a detector can be technically right while the company is an unsuitable target, and an appropriate account can have too little public DNS evidence for a technology-specific opener.
At the handoff, each row should answer: who is the account, why might this service fit, which observations support that hypothesis, how recent are they, and what must be checked manually? Keep those answers in explicit columns instead of burying them in an unstructured note. It makes coaching and later funnel analysis possible.
Turn infrastructure signals into a qualified prospect list.
Analyze domains you already know, or find new companies that match the technologies you target. Public DNS observations support discovery, not proof of an internal security weakness.