01 / Context
High demand, delivered by hand
The client is a mid-sized corporate services business based in the UAE, specialising in company formation and related advisory services. A significant business unit focused on setting up companies in the a major UAE free zone, a high-volume but operationally demanding offering.
Demand for free-zone setups was strong, but the underlying delivery model relied heavily on manual effort. Customers made initial contact via the website, after which internal staff manually qualified requirements, explained eligibility rules, and prepared costed quotes. Much of this work was repetitive and predictable, yet still required experienced staff to execute. As volumes rose, the unit suffered from high website bounce rates and a sharp drop-off after first contact. Each lost customer represented roughly AED 15,000 a year in fees, materially affecting revenue and margins.
02 / The Challenge
Skilled people tied up in routine qualification
The core challenge was structural rather than tactical. Five full-time staff were tied up in administrative qualification and onboarding work, while consultants spent significant time delivering standardised advice. This eroded margins and created inconsistency in customer guidance.
Incorrect or inconsistent advice led to change fees, lower retention, and reputational risk. Slow response times contributed to drop-off, and scaling the unit further would have required more headcount simply to hold service levels. The client needed to improve the consistency, speed, and accuracy of advice while dramatically reducing human involvement in routine cases.
03 / The Approach
A GenAI agent that qualifies and quotes
We designed and delivered a GenAI-powered customer qualification and quoting solution, centred on a conversational chat experience: a custom chat widget embedded on the client website, an agent that qualifies customers in natural language, and a structured knowledge base encoding the free-zone rules, eligibility criteria, pricing logic, and constraints.
Once qualification was complete, the agent generated fully costed, accurate quotes automatically, and escalated to a human only when edge cases or non-standard requirements were detected. It delivered consistent, policy-aligned advice every time, responding instantly and removing friction from the early journey. The initial deployment focused on a single free zone, with the architecture designed to support future rollout across multiple free zones.