Case study
How BCIT tested conversational searchagainst traditional site search
Early results showed stronger revenue per user and conversion among people who engaged with Airgentic.
Compared with users of traditional site search during the first two months of operation.
2.3×
higher revenue per user
3.4×
higher overall conversion rate
<5%
of traditional search volume
Airgentic usage in the period
A fair test against a search experience that already worked
British Columbia Institute of Technology (BCIT) deployed Airgentic to explore a simple but important question: can conversational search become a more effective service and conversion channel than traditional site search in a higher education environment?
The challengeFinding the right path through a complex website
BCIT has long understood the value of search in helping students find what they need. Like many higher education institutions, it supports multiple audiences through a large and complex website. Despite significant ongoing work to create clear and meaningful digital experiences, some users can still find higher education websites difficult to navigate.
- Prospective students may understand the outcome they want without knowing the right program, pathway, requirement, fee structure, intake or support option.
- If they cannot quickly understand their choices and next steps, momentum can be lost.
- Service teams may receive enquiries that could otherwise be resolved through self-service.
The institution needed a way to help people move forward from a sentence, not a search term they might not know.
Sitewide conversational search, grounded in approved content
Airgentic was launched sitewide at BCIT in mid-January 2026, with indexing across the institution's approved content. The platform gives users grounded answers and helps them move forward without requiring them to search through pages and PDFs on their own.
What the platform gives users and the institution
Grounded answers
Purpose-optimised retrieval for complex institutional content.
PDF understanding
The ability to interpret information in tables, diagrams and images.
Human oversight
Tools to review, override and take over conversations where needed.
Conversational intelligence
Insight into user questions, areas of confusion and content opportunities.
Why BCIT chose this approach
The BCIT team wanted a rigorous test rather than a hype-driven experiment. BCIT already had a working search experience and understood its value. The real question was whether conversational search could produce a measurable uplift over that existing baseline.
The team used existing analytics and course-sales data to compare the performance of users who engaged with conversational search against users of traditional site search. This moved the project beyond a usability exercise and created a practical comparison of service and conversion outcomes.
Promising early indicators, not a ceiling
After roughly two months of operation, BCIT observed that users engaging with conversational search had 2.3 times higher revenue per user and a 3.4 times higher overall conversion rate than users of traditional site search.
The comparison was meaningful because BCIT already knew traditional site search performed well. Airgentic usage volumes were also less than 5% of traditional search volume during the period, making these promising early indicators rather than a final view of the platform's potential at broader adoption.
2.3× revenue per user
Observed among people who engaged with conversational search, compared with traditional site search users.
3.4× conversion rate
Higher overall conversion among the same conversational-search group.
Under 5% of search volume
Airgentic usage during the period — early adoption, not full-scale replacement of site search.
A strong baseline
Traditional site search already performed well, so this is an uplift on an established channel.
Measurement note
These figures compare observed user groups during the early deployment. They are not presented as a calculation of revenue earned or costs saved directly as a result of Airgentic.
Why the results matter
The BCIT project suggests that conversational AI can do more than answer questions. When designed well, it can help users understand complex options, reduce friction at important moments and guide people toward a suitable next step with greater confidence.
For institutions, the same interaction can also reveal what people are asking, where confusion exists and where content or services could be improved. This combination of service improvement, conversion performance and conversational insight is central to Airgentic's approach for higher education.
Efficiency and implementationIntelligent retrieval, not a sledgehammer
BCIT also valued Airgentic's technical approach. One of the criteria that mattered to the team was that the platform performed the task intelligently, without relying on excessive processing power or a 'sledgehammer' approach for a precise task on a relatively small data set.
- The work is retrieval-led rather than brute-force processing of the whole corpus.
- The deployment can be controlled, efficient and sustainable.
- That fit matters for institutions that care not only whether AI works, but how it is run.
Airgentic's retrieval-led design was seen as a positive fit for that requirement.
What comes next
The deployment creates a foundation for more specialised conversational journeys, including course and program enquiries and other targeted support pathways. It also gives BCIT a richer intelligence layer for identifying patterns in demand, content gaps and service friction across the organisation.
As adoption develops, the opportunity is to extend from sitewide guidance into more targeted student acquisition, service and support experiences while maintaining institutional oversight and grounding responses in approved content.
A grounded, governable answer layer for complex public-facing environments
BCIT's early experience indicates that conversational search can deliver a different level of service and conversion performance from traditional site search. In the first two months, users who engaged with Airgentic showed 2.3 times higher revenue per user and a 3.4 times higher overall conversion rate than site search users.
For higher education institutions, the project is an early example of how a website can become more than a repository of information. It can become an active service channel that helps people reach the right answer, program, service or next step.