Identify whether the team needs knowledge or time
A capable Slate team can still struggle to make progress. Tristan Deveney describes experienced staff who know how to build but spend their days answering questions, repairing inherited complexity and completing the next cycle’s work. Another training session may not address the problem if the missing resource is uninterrupted time.
Separate the work that requires institutional knowledge from the work someone else could support. A captain may need to retain decisions involving campus relationships or local processes while another person handles a recurring technical question. An agreed escalation route can protect that focus without leaving colleagues stuck.
Choose one improvement that keeps being postponed. Name the person who needs time for it, the interruptions that regularly displace it and the support that would make a protected work block possible. Make the tradeoff visible before adding another commitment.
In the recording · 14:21 ↗Build a version people can react to
Tristan’s description of agile development centers on a working version that users can try. Early feedback reveals needs that a long specification may miss, and the next iteration can respond to what actually happens.
Define the smallest complete task the first version should support. Give a staff member a realistic example, watch where they hesitate and record what they expected to happen. Distinguish a missing essential behavior from an attractive extra. Agree on the next change together rather than treating every suggestion as immediate scope.
A prototype still needs boundaries. Decide which data and permissions it uses, what can be changed and who reviews the result. Keep the first test small enough to understand, then carry the lessons into the next version.
In the recording · 09:56 ↗Review the journey before the form
The inquiry-form research begins with a prospective student’s question: how do I ask this college for more information? Tristan starts at institutional homepages, counts the clicks needed to reach an inquiry form and then examines Slate forms more closely. This connects navigation decisions with the experience inside the form.
Try the same walkthrough without using a saved direct link. Record the labels you follow, the places you expect to find the form and any step that sends you back. Repeat on a phone and ask someone unfamiliar with the site to try it.
The research describes observed website practices, not a measured conversion effect. Use those observations to choose a question to test locally. A hard-to-find link deserves investigation; the click count alone does not tell you how many inquiries were lost.
In the recording · 19:00 ↗Bring the website and form owners into the same review
Tristan points to mobile branding as one place where the experience can differ from what a marketing team expects. A form may look familiar in one browser and show a different template on another device. Annual prompt updates can leave that wider journey unexamined.
Review the landing page, form, mobile presentation and confirmation together. Check that instructions remain readable, required fields have a clear purpose and the next step is understandable. Share the actual device behavior with both the website owner and the Slate team.
The conversation also notes emerging AI-related consent language. Have the appropriate policy owners review the tools actually in use and the permission being requested. Copying another institution’s wording does not establish that it fits your communication process.
In the recording · 23:07 ↗Let the next question become a learning opportunity
When time allows, Tristan coaches a colleague through the problem instead of only returning the fix. Explaining a thought process can expose an assumption, suggest a useful log to inspect or help someone reproduce an unexpected result.
Ask the colleague to describe what they expected, what happened and the smallest example that demonstrates the difference. Work through the evidence together. Let them explain why the proposed change addresses the cause, then test the result.
Save the useful pattern afterward: the symptom, the checks that mattered and the limitation of the fix. Urgent incidents may need a faster response, but the learning can follow once service is stable. A short, searchable example gives the next person somewhere to begin.
In the recording · 24:45 ↗Use an internal portal to coordinate shared work
Tristan describes internal portals for matching and scheduling work previously handled through spreadsheet exports and reimports. One example separates firms from their internship slots and links slots to candidates. Other examples coordinate workshop preferences or interview availability.
Start by defining the records and relationships the workflow needs. Identify who can see a record, who can change an assignment and which updates should trigger a message. Keep selection decisions with the responsible people; the portal organizes their work.
Test competing edits, capacity limits and stale information before relying on a shared view. Decide how an exception is corrected and how staff see the current status. Involve the team during the build so the finished workflow becomes something they can explain, maintain and adapt.
In the recording · 30:08 ↗