Turning a program catalog into a conversation.
Prospective students know they want a healthcare career; far fewer know which of a hundred-plus offerings gets them there. For a health sciences university we built an AI program advisor that meets prospects mid-question, recommends the programs that actually fit, and walks every conversation to a next step admissions can see.
The challenge
Prospective graduate students often know they want a healthcare career but not which path is right for them. At this institution that uncertainty meets real scale: more than 100 academic offerings, including 35 degree and certificate programs spread across three schools of nursing, rehabilitation sciences, and healthcare leadership, plus continuing-education courses and a wide range of prerequisites. The prior experience asked overwhelmed prospects to diagnose their own fit from a complex catalog.
That put the burden on exactly the person least equipped to carry it. A catalog is organized around what the institution offers; a prospect’s question is organized around their own life. “I’m a nurse, what’s my next credential?” “I’ve never worked in healthcare, where do I start?” Every gap between those two framings was a place to stall: an unfound prerequisite, an unclear deadline, a program page that assumed you already knew it was the right one.
The approach
The answer wasn’t a better catalog; it was a translation layer between the catalog and the person. We designed a conversational advisor that starts from the prospect’s question, maps it across everything the institution offers, and treats every answer as one step in a path toward applying, not a dead end.
Start from the question, not the catalog
The advisor meets prospects wherever they land on the site and opens with guided “kickstarter” prompts like “What program is right for me?”, so the first move costs nothing. From there it classifies what a prospect actually wants and routes the conversation accordingly: entering nursing, advancing an existing credential, comparing clinical paths. The taxonomy stays the institution’s problem, not the prospect’s.
Answer like an advisor, not a search box
Once it understands intent, the advisor maps it across all offerings and recommends the programs that best fit, then keeps answering in context. The specific requirements, curriculum, clinical experiences, costs, and deadlines relevant to that prospect surface inside the conversation, instead of being scattered across the program pages where the same questions used to go unanswered.
End every conversation with a next step
As a conversation progresses the advisor points to the next logical move: attend an event, schedule time with admissions, start an application. Each of those is tracked, so curiosity converts into a step the admissions team can see and act on. That is the difference between a chatbot that answers questions and an advisor that drives acquisition.
“What program is right for me?”
What changed
The advisor went live in May 2026, and the early conversation data reads like a map of what prospects had needed all along. More than half of sessions are program exploration, literally “what program is right for me,” and another 35% are admissions-requirements questions: deadlines, dates, prerequisites. A June sample of 110 tracked conversations told the same story in finer grain, with prerequisites and application deadlines the most-asked topics, and nursing the most-asked-about field, both by people entering it and by nurses advancing an existing degree.
In other words: the questions the catalog structure had been quietly failing to answer are now the advisor’s main job, asked in the prospect’s own words and answered in context. And because every conversation points somewhere, the interest is no longer anonymous: in the first weeks, prospects submitted 22 connect-with-an-advisor forms straight from their conversations.
The early returns
Weeks in, the numbers are early and directional, but they all point the same way: prospects would rather ask than browse, and their conversations end in steps admissions can act on.
- 0
- Connect-with-an-advisor forms submitted in the first weeks
- 0%+
- Sessions exploring which program fits, the question the catalog couldn’t answer
- ~0
- Conversations a day since launch
What I’d carry forward
- 01
Meet the question, not the taxonomy. Prospects don’t browse a hundred offerings; they ask one thing about their own life. Guided prompts make the first question free, and intent, not the org chart, should route what happens next.
- 02
Treat the conversation log as research. The session mix, half program fit and another 35% requirements and deadlines, is the clearest voice-of-prospect data yet on what the site fails to answer.
- 03
An AI answer should end in a human next step. Advisor connections, events, and application starts are what make a conversational tool an acquisition channel rather than a novelty; instrument them from day one.
Have a system that has outgrown its structure?
This is the kind of problem I like best. Tell me where it stands and I’ll tell you how I’d approach it.