In November, WGU Labs went to Houston City College expecting to do some convincing. We thought the advising team would be skeptical of the AI tool we designed to absorb some of their work, but the opposite was true. The advisors were open, even eager. They were ready for something to take over the transactional parts of their jobs.
The team had one condition: making sure the tool could meet the logistical, navigational needs of real HCC students. HCC advisors wanted more AI support; our team needed to figure out the right kind of “more.”
After working through exercises to identify common HCC student personas, common challenges, and examples of the questions they typically bring to their advisors, precision, personalization, and preparation emerged as top design priorities. Many HCC students have foundational questions about navigating college, which meant we couldn’t send them to a webpage full of links to sort through on their own. The AI agent would need to surface personalized answers for each student: a specific professor’s name, a direct contact, the exact form, the exact deadline.
Many students also need help navigating critical decisions in their program, especially what courses to register for and how to sequence their program to stay on pace to graduate. Advisors are there to help students with these decisions but their time is limited. When students show up prepared to talk about these decisions, the advisor can support them better. The AI agent would need to surface these critical decisions for students and help them think through them, and prepare them to have a productive discussion with their advisors.
Essentially, the advisors shared that foundational questions about office hours, building locations, and transcript requests constitute transactional work. These tasks are important to students but can crowd out the more complex, transformational work like providing career direction, degree planning, and advocating for students in academic or financial crises. Fortunately, a new Anthropic study finds that transactional tasks are the very tasks AI is suited for. HCC advisors want to leverage that.
The way the team sees it, the work of SwoopChat — named for the HCC mascot — is less to replace them and more to prepare students for productive advising appointments that make the best use of everyone’s time. If SwoopChat can answer preliminary questions, students can arrive at their appointments ready to ask better ones. With HCC advising caseloads between 400 and 800 students, that kind of preparation can have real impact.
The Houston onsite also helped close the distance between the SwoopChat user we’d imagined and the real HCC student who would be using the tool. The advising team reminded us of something we knew from our prior research: 80% of rising-talent learners are deskless workers, many HCC students included. We knew HCC’s student population skewed toward first-gen learners who often have limited time and resources. Even so, we had been designing SwoopChat for a student with a laptop, a desk, and a quiet hour to spare.
During the persona exercises, we learned that HCC students do more than we realized on mobile: registering for classes, submitting forms, completing coursework. The onsite helped us see that mobile is more than a feature layer of SwoopChat. For many HCC students, it’s the full product. This was an important outcome of our visit and had a significant impact on the design direction we took from then on.
The button choices for our chat screen were also directly informed by conversations with the advising team. Need help paying for classes? How many classes do I have left? Who is my advisor? Prep for my advisor meeting? Not sure which career you want to start? When students open SwoopChat, these prompts help guide them to some of the most important types of support that SwoopChat can provide.
SwoopChat’s next step is a rich user experience study for summer, 2026 in which students try out the tool and give us their feedback on what they like, don’t like, or think is missing. We are aiming for a wider pilot in fall 2026. That quick timeline is intentional. WGU Labs aims for speed to insight: a development rhythm that moves quickly from prototype to real users to evaluation.
It won’t be our first pilot with an AI student support agent. We ran a similar pilot earlier this year with WGU students, using an agent called Stu. The HCC findings will be different. This is a different institution, a different student population, and a different set of questions. We are committed to listening closely.
