How a multi-branch visa consultancy cut pre-call workload by 65% with AI

A multi-branch visa consultancy came to us with a familiar problem: consultants were spending a large share of every day answering the same handful of eligibility and documentation questions, before a single client ever got to a paid consultation. The volume was growing with the business, and hiring more consultants just to triage questions wasn't a scalable answer — especially in a field where a wrong answer has real consequences.

The brief was narrow on purpose. They didn't want an AI that gives immigration advice. They wanted the repetitive, factual front door — 'am I eligible to apply,' 'what documents do I need,' 'how long does this visa type take' — handled automatically, with anything requiring judgment routed straight to a consultant.

We started with discovery, not a build. Before writing anything, we sat with the branch teams and mapped which questions were genuinely repetitive and rules-based versus which ones depended on a client's specific situation and needed a licensed consultant's judgment. That line — not the AI's capability — is what defined the scope.

The system we built answers eligibility and documentation questions instantly, pulling from the consultancy's own visa category rules and requirements rather than general knowledge, so answers reflect their current process across every branch. Anything that touches individual circumstances, edge cases, or actual advice gets flagged and handed to a consultant with the conversation history attached, so nothing gets asked twice.

Within the first few months, 65% of routine eligibility questions were resolved before a client ever booked a call. Consultants went into consultations with clients who already understood their documentation requirements, which shortened the calls that did happen and let the team take on more clients without adding headcount.

The part that mattered most to the client wasn't the automation — it was the boundary around it. Every routed handoff includes the reason it was escalated, so consultants can see at a glance whether it's a straightforward volume issue or a case that needs real judgment. That audit trail also made it easy for them to review accuracy and adjust the rules as visa requirements changed.

If you're running a consultancy or advisory business where clients ask the same qualifying questions before every engagement, the lesson here isn't 'add a chatbot.' It's that the highest-leverage use of AI is usually the boring, repetitive 20% at the front of your process — freeing your experts to spend their time on the 80% that actually needs them.

That's the same question we'd ask about your process before recommending anything: which parts are genuinely repetitive, and which parts need a human? Worth a conversation before assuming the answer is a bigger build than you need.