Put the context where the work happens
A visitor meets a tour guide. A counselor visits a school. A campus guest chooses an information session with its own check-in location. Each experience creates a relationship that the next message ought to remember. When those connections live only in someone's notes, personalization becomes another round of copying, checking and sending.
RARE, the Related Automation and Relationship Engine, starts by attaching that context to the event in Slate. The recording works through three uses: a tour-guide autoresponder, a post-visit message to a school contact, and event-specific instructions inside a shared communication template. The useful pattern is to identify the relationship, retrieve the details that belong to it, and let those details select appropriate content.
In the recording · 01:15 ↗Make the tour-guide message earn its introduction
The first build begins with event form responses. Attendance matters: a message that says it was good to meet someone should not go to a person who never attended. The recipient logic also checks that a tour-guide Related Record exists. Those two checks give the message a defensible starting point.
Getting to the guide is only part of the join. In this demonstration, the guide's individual field values need their own joins before they can become exports. A dataset filter distinguishes the guide from other Related Records attached to the same event. Without that distinction, a parking record and a person can occupy the same conceptual bucket even though the message needs very different information.
Once the guide's major and the visitor's interest are available, a comparison produces a consistent yes/no flag. Liquid then chooses between two paragraphs. Start with two fictional visitors, one who matches and one who does not. Confirm both branches, then check a missing value separately. A blank interest should never quietly become a confident claim about what someone wants to study.
In the recording · 08:15 ↗Keep the school contact attached to the right visit
The second build moves to organization contacts. A contact can be associated with a visit, but that association deserves a cross-check against the event's organization. People change schools, and a familiar name can be selected from the wrong record. The demonstration approaches the visit through both the related contact and the organization/location relationship.
Recency is another part of accuracy. The recording discusses delayed imports and old events that might otherwise produce an oddly timed follow-up. Decide which visit dates qualify and how the most recent relevant event is selected before writing the greeting. A technically valid join can still describe a conversation that happened months ago.
The staff member associated with the event can supply the key for an introductory content block. A shared naming pattern makes those blocks easier to find and maintain. The recording sketches this mechanism rather than building every counselor's finished message, so the content library, ownership and fallback still need deliberate preparation.
In the recording · 22:44 ↗Let event details vary without multiplying templates
The third build tackles the familiar problem of a shared template with different parking or check-in instructions each week. Instead of rewriting the template before each event, an event-details dataset stores the variation. A structured type and a recognizable record name make the right detail easier to attach.
An umbrella registration can include a separate information session. The communication needs to reach that session's Related Records, not simply whichever event record is closest in the query. The demonstration follows the person's registration to the relevant session and checks that its date matches the umbrella visit. Comparing dates rather than identical date-times matters when the umbrella and session begin at different hours.
Check-in is the worked example. Parking and more elaborate routes are extensions, not completed branches in this recording. Define a sensible default when no override exists, and test conflicting or duplicate detail records. When instructions change, the next generated message can use the updated relationship; an email already in an inbox will not change.
In the recording · 41:27 ↗Test the relationship before polishing the paragraph
The live build exposes a small dependency with a large effect: record names must exist before someone can reliably find and attach those records. Missing naming rules lead to a workaround on screen. That is a useful reminder to test the authoring process as well as the recipient query.
Take one of the three patterns back to your workbench. Draw the path from recipient to event to Related Record, then name the rule that prevents the wrong relationship from qualifying. Test a correct record, a missing record and a plausible wrong record. Only then refine the wording. The personal touch becomes more trustworthy when the underlying relationship is something the system can explain.
In the recording · 47:01 ↗