Amazon SES Bot Click Detection: What It Means for Sales Follow-Ups

On August 7, Amazon SES added a field that should make sales teams look again at their click-triggered workflows. Open and Click event notifications now include isBotEvent, which indicates whether an interaction was likely automated. Its possible values are Likely and Unlikely. For customers already publishing these events through a configuration set event destination, the field is included automatically. AWS announced the change here.

The immediate question is what your software does with that information. If an email click adds points to a lead score, alerts a salesperson or starts another message, a machine-generated interaction can influence the next thing a person receives. Adding a classification to an event only helps if the workflow actually reads it.

A Click Records an Action, With Limited Context

SES click tracking works by sending a request through a tracking URL before redirecting it to the destination. That gives the sender an observable event. It does not, by itself, explain who initiated the request or what they intended to do. AWS also documents ways recipient-provider privacy settings and caching can affect tracking. Its metrics FAQ describes these mechanics and limitations.

That distinction matters when a sales rule assigns meaning to a request. Visiting a pricing link might justify looking more closely at an account. It does not establish that the recipient read the pricing, discussed a purchase or wants another email. Even a correctly identified human click cannot settle those questions.

The new field makes one uncertainty easier to examine. AWS describes it as a likelihood, and the event schema places it inside both the Open and Click objects. Unlikely means the event was classified as unlikely to be automated. It is not a verified statement of buying intent.

Review What Happens After the Click

Start with the rules that change a recipient's experience. Does one click shorten a follow-up interval? Does it move someone into a more persistent sequence? Does it create a task saying the prospect is interested? Each rule contains an assumption worth making explicit.

For example, take a workflow that sends a follow-up whenever a recipient visits a product link. A better version would retain the original event, record its bot classification and keep likely automated activity from independently triggering that message. The salesperson could still see the event while judging it alongside replies, previous conversations and other relevant context.

Handle missing classifications explicitly, too. Historical records and data from other providers may not contain this SES field. If an integration silently converts a missing value into a human click, the report will look more certain than its evidence allows. Keep that uncertainty visible.

Your campaign software still needs rules for scoring leads and deciding whether to send a follow-up.

Compare Results Without Losing the Original Record

Before changing your headline click metric, keep a version of the original count. Then show how much activity was classified as likely automated and how much remains. Record when that definition changed. Otherwise, a cleaner measurement can look like a sudden campaign decline to someone comparing this month's results with last month's.

Apply the same discipline to business outcomes. A reply, an agreed meeting and a completed purchase each tell you something different. Track the outcome the campaign is meant to produce, and use clicks as supporting evidence rather than allowing them to stand in for it.

At Mailrun, we want the sending setup and the campaign decisions to be evaluated together. Before the next campaign goes out, find the rule that turns a click into another message. Check the evidence it uses and the message it produces. That is a concrete place to put AWS's new signal to work if your sending stack uses SES.