For the American University of Antigua (AUA), the challenge was straightforward: too many prospective students to reach and not enough hours in the day.
AUA receives roughly 9,000 leads each month, supported by a team of about 12 admissions counselors. Reaching those prospects requires counselors to make hundreds of calls per day, many ending in voicemail or going unanswered.
The university saw an opportunity to use AI differently. Instead of simply answering questions, AUA wanted AI agents that could engage prospective students, complete parts of the admissions workflow, and connect qualified prospects with its team.
The result was an agentic AI deployment that moved from initial conversations to contract in roughly 45 days, with AUA's first pilot going live in about four weeks.
Moving Beyond the Traditional Chatbot
AUA already had experience with a traditional website chatbot. But leadership wanted something capable of doing more than providing information.
Working with Druid AI, AUA initially focused on two admissions use cases.
The first was a student-facing AI agent capable of answering questions about admissions requirements, financial aid, leave-of-absence policies, and other common topics.
The second was an AI voice agent designed to address one of the admissions team's biggest time drains: outbound calling.
Instead of counselors repeatedly dialing prospects who may not answer, the AI agent can initiate the conversation, determine whether the prospective student is ready to speak with someone, and then make a live transfer to the admissions team.
That allows counselors to spend more time doing what they are there to do: advising and engaging prospective students.
Channel mattered as much as capability. AUA's student base includes a large international population, and about 80% of them are on WhatsApp. The agent sees engagement rates of 60% to 65% there. "That's big," said Sheikh Sadiq, Executive Vice President of Strategy and COO at Manipal Education Americas.
Start With a Real Problem
One reason AUA was able to move quickly was its approach to the project.
Rather than beginning with the question, "How can we use AI?", the university started with a specific operational problem.
Admissions and IT worked together to define what needed to improve and what success should look like. Executive leadership was brought into the process early, including AUA President Dr. Peter Bell, who sat in on a live product demo and watched Druid's team run real calls. By the end, his question wasn't whether to move forward. It was, "How many agents are we getting?"
The technology also had to be demonstrated against AUA's actual requirements. Sheikh Sadiq said the process was less about buying a predefined solution and more about understanding what AUA wanted to accomplish.
In our customers' words
"The reason I picked Druid is because their agent is a ready-made agent, it's already there," Sheikh Sadiq said. "You give it commands, feed it data, and tell it what responses you want. That's the advantage: you don't have to do a lot of development work."
That alignment helped AUA move from evaluation to implementation without months of custom development.
Build, Test, Then Scale
Speed did not mean simply turning on an AI agent and hoping for the best.
AUA's admissions team, IT organization, and Druid implementation team tested responses, tightened the knowledge available to the agents, and established boundaries around where information could come from.
The university took the same approach with outbound voice.
Rather than immediately launching thousands of calls, AUA began with 50, then scaled to 100, then 125. It's now running about 200 calls a day.
That iterative model is important for higher education institutions evaluating AI: start with a defined use case, establish governance, test it with real users, and expand from there.
The deployment also surfaced an unexpected benefit. Because the AI agent relied on information from AUA's website, testing helped the university identify outdated content and broken links that needed to be corrected.
What Comes Next
Admissions is only the beginning of AUA's AI roadmap.
AUA's president has set up an AI committee to guide the university's next phase of work, alongside an academic committee focused on AI in the classroom.
The university is now exploring additional agents for document collection, financial aid, and academic support. One planned use case would help the admissions team follow up with students about transcripts, recommendation letters, and other outstanding documentation. AUA expects that agent to be live within 30 to 45 days.
AUA is also exploring how AI could help students identify academic challenges earlier in the semester, giving both students and faculty more time to act.
Find a meaningful problem. Bring the right teams together. Establish the appropriate guardrails. Prove the value. Then build from there.
For AUA, that process happened in weeks, not semesters.
In our customers' words
"AI is mainstream. Don't get left behind, pick the right partner," Sheikh Sadiq said during the webinar. "Make sure you know what you want out of AI. Do not get into a lot of development work, that would be wrong, because you'll be developing forever and ever and ever. Take small problems, find solutions. Druid was the right partner for us, and as you implement these little projects and keep building on it, that's the way to go. Definitely don't wait. You'll miss the boat."
Watch the full conversation: Weeks, Not Semesters: How AUA Deployed Agentic AI
