A recorded click is evidence. It is not automatically proof of a person.
Security scanners, mailbox protections, link checkers, privacy systems, corporate gateways, and other automated tools can open or click email links before a real recipient has touched the message.
Advanced Subscriber Intelligence evaluates thirty-eight signals across a 72-hour evidence window to estimate whether recorded engagement looks automated, uncertain, probably human, or strongly supported by human evidence.
ASI assigns a continuous human-confidence score from 0% to 92%. It preserves the reason evidence beneath that score rather than turning one click into a cheerful but unsupported sales lead.
72-Hour Evidence Window
0% to 92% Confidence
Reason Evidence Preserved
Clicks should be interpreted from the evidence around them.
The same URL can be triggered by a security scanner, an automated prefetch system, a recipient checking the message on a phone, or a person deliberately navigating through the campaign. The event alone does not reveal which one happened.
Why ordinary click totals can mislead the business
Automated activity can make a campaign look busier, more successful, and more commercially useful than the underlying evidence supports.
Inflated performance
Scanner activity can increase total clicks and click-through rates without representing additional human interest.
Poor follow-up priorities
Sales and account teams may waste time contacting recipients whose only recorded activity came from automated security systems.
Wrong campaign learning
Subjects, offers, links, audiences, and content can be judged incorrectly when automated behaviour is treated as customer intent.
Unreliable client reporting
A large raw total may look impressive in a presentation while failing to explain who engaged, why the event looks human, or how certain the system actually is.
Thirty-eight signals. One evolving interpretation.
ASI combines multiple weak and strong signals rather than trusting one browser string, one timestamp, or one click event to decide the answer.
Timing evidence
How soon the event happened after send or delivery, whether the sequence looks physically plausible, and whether several links were triggered too quickly for normal reading.
Navigation evidence
Browser navigation, page movement, later actions, click sequence, and related behaviour can strengthen the interpretation after timing supports human probability.
Device and browser evidence
Operating system, browser, device, user agent, network behaviour, and consistency across events contribute context without being treated as proof on their own.
Scanner and automation evidence
Known machine patterns, corporate gateways, security tools, prefetch behaviour, repeated signatures, and link-checking sequences can lower confidence.
Honeypot evidence
Interaction with hidden or non-human-facing links is strong evidence of automated activity and is preserved as part of the reason trail.
Later-event evidence
A suspicious early click may later gain stronger human evidence. ASI can revise the interpretation as new activity arrives within the 72-hour window.
A browser label cannot rescue an impossible click.
ASI does not award strong human confidence merely because an event claims to come from Apple, Chrome, Safari, Windows, or another familiar environment. The timing and behaviour must first support a plausible human action.
Very early click
A click occurring almost immediately after send or delivery is treated as suspicious even when the user agent looks like a normal browser.
Rapid multi-link sequence
Several links triggered within an unrealistically short period can indicate automated link checking rather than normal recipient navigation.
Plausible timing plus browser support
Browser, device, and navigation evidence can strengthen confidence only after the timing evidence supports human probability.
Later human rescue
A recipient may first generate scanner-like evidence and later produce a more plausible human sequence. ASI preserves both parts of the story.
A continuous score, not a magic on-off switch.
ASI keeps the full 0% to 92% confidence score. Common operational milestones help operators understand the strength of the evidence without pretending the boundary between bot and human is perfectly sharp.
Automated, suspicious, or insufficient
The available evidence is dominated by automation signals, implausible timing, weak context, or too little information to support likely human engagement.
Human probability begins
Timing and behaviour begin to support a likely human interpretation, while uncertainty and contradictory evidence remain visible.
Strong human evidence
Plausible timing is supported by stronger browser, navigation, device, sequence, or later-event evidence.
Highest current operational confidence
The strongest available evidence supports likely human intent. ASI still describes this as confidence, not absolute certainty.
The first event is not always the final answer.
ASI continues evaluating relevant engagement evidence for 72 hours so later actions can strengthen, weaken, or clarify the original interpretation.
Initial event
ASI records the click, timing, link, recipient, campaign, user agent, provider context, network evidence, and immediate classification signals.
Related activity
Later opens, clicks, navigation, repeated behaviour, different devices, or new automation evidence can change the weight of the original event.
Reclassification
The confidence score and reason codes can evolve while the historical evidence remains preserved rather than being silently overwritten.
Final reporting context
Reports Builder, Event Stream, Recipient Drilldown, CSVs, and PDFs consume the current scoped interpretation while retaining the evidence beneath it.
The score is useful because the reason remains visible.
ASI does not ask the operator to trust a confidence number floating in space. The reporting workflow preserves the evidence used to reach the classification.
Reason codes
Timing gates, browser navigation, scanner evidence, honeypot activity, device support, automation patterns, and later-human evidence remain explainable.
Event history
The operator can inspect the related sequence rather than relying only on the latest classification label.
Recipient context
Recipient Drilldown connects the confidence result to the campaign, links, delivery, device, timing, and other permitted subscriber evidence.
Scoped exports
Reports and CSVs can be filtered by confidence threshold, link, campaign, recipient, time, and selected fields without rebuilding a separate click truth.
The goal is not fewer clicks. It is a more useful call list.
ASI keeps the raw evidence, filters automated noise, and lets teams choose the confidence threshold appropriate to the action they plan to take.
Preserve raw clicks
Do not destroy the original activity merely because it looks automated. Raw evidence remains useful for provider, scanner, security, and campaign analysis.
Apply human-confidence evidence
ASI scores the event using timing, navigation, device, scanner, honeypot, sequence, and later-event evidence.
Choose the operational threshold
A campaign review may use a broader threshold, while a sales call list may require stronger confidence before somebody spends time contacting the recipient.
Export the evidence
Build a scoped recipient or link export containing the permitted identity, company, role, phone, website, confidence, reason, and campaign evidence needed for follow-up.
Up to 92% accurate operational data, with uncertainty left intact.
ASI reporting is designed to provide up to 92% accurate operational data. The confidence model improves the interpretation of click evidence, but it does not see inside a person’s mind or control the private behaviour of every mailbox provider and security system.
Confidence is not certainty
A 92% score means the available evidence strongly supports likely human intent. It does not prove the recipient’s identity, motivation, or commercial intention beyond doubt.
Automation keeps changing
Mailbox protections, scanners, privacy systems, browsers, and network tools evolve. Classification rules and evidence interpretation must remain under review.
No invented engagement
ASI will not upgrade a weak event merely because a higher confidence total would make the campaign look better.
Gold Standard remains under study
The dormant Gold Standard confirmation concept remains disabled while ASI studies whether additional confirmation can improve evidence without creating new scanner or recipient-friction problems.
Any team using click data to decide what happens next.
Human-confidence reporting becomes commercially important when campaign data drives sales calls, segmentation, client reporting, campaign optimisation, account follow-up, or sender strategy.
Sales teams
Reduce the chance of following up with recipients whose activity is supported only by machine-generated clicks.
Campaign operators
Understand which links and recipients show stronger human evidence before deciding how the next campaign should change.
Agencies and client teams
Explain the difference between raw activity and likely human engagement without presenting confidence as absolute fact.
Reporting reviewers
Inspect the events, reasons, thresholds, and recipient context behind an engagement figure before accepting it.
Explore the surrounding ASI rails
Continue into Campaign Reports & Insights, Advanced Data Exports, Sending Engine & Queue Control, List Health & Validation, Subscriber Validation Pipeline, and Email Compliance.
Stop treating every machine click as customer intent.
Request a private Sender Review to examine how your current platform records clicks, handles scanners and privacy systems, supports recipient drilldown, exports follow-up lists, and explains the confidence behind the engagement numbers it displays.