AI Virtual Assistant Answering Prospective Students 24/7

Post University is a private, for-profit university in Waterbury, Connecticut, USA, known for being one of the first institutions to offer full degrees online. They have an internal department responsible for data management and business intelligence.

Prospective students came to the website with questions about programs and how to apply, whenever it suited them. The project’s first goal was to answer them while they were still interested.

A website visitor became a lead only after leaving contact details, and the university wanted more of its visitors to reach admissions advisors that way.

Earlier automation on the website followed fixed logic that only a developer could change, so the university’s own team couldn’t make even small improvements.

We built an AI assistant that talks with prospective students on the university website, by text or voice. A visitor says what they are looking for, and the assistant suggests matching programs and answers questions about studying at the university. It answers from the university’s website and conversation scripts, so its answers stay in line with what the university tells applicants. The program catalog behind its suggestions updates automatically. 

We connected the assistant to the university’s CRM. It asks visitors for their name, email, and phone number. When a chat ends, its transcript and a short summary go to the CRM, so an advisor can see what the person asked before calling.

Guardrails check each answer before the visitor sees it. When they block an answer, the assistant writes a new one, up to three times, and otherwise declines politely.

Less than six months after the project started, the assistant went live on the university website, giving prospective students program suggestions and answers at any hour.

Advisors can open the first call knowing what a prospective student asked and which programs came up.

The project continued with an AI assistant for the university’s admissions advisors, which answers their questions from the university’s own manuals and scripts. Questions about the university’s numbers go to Ava, a separate assistant built on its data platform.