Aubergine Solutions makes complexity behave

Someone in Aubergine Solutions’ leadership wanted a number: how many hours had the design team spent on a particular project last month? Insifi, the company’s internal AI assistant, supplied one. It sounded certain. When checked against the timesheets, it was wrong. The tool had mixed up similarly named projects and included people outside the team. A perfectly ordinary office question had uncovered a rather consequential design problem.

Insifi was connected to Nexus, Aubergine’s system for employee information, timesheets and project allocation. Early tests had looked promising. People used it; complaints were scarce. The mistaken answer changed the evidence. The team assembled roughly 200 to 300 questions, scored them manually and used the failures to locate troublesome joins, naming conventions and phrasing. The lesson is useful well beyond this company: a quiet support channel can coexist with a busy, confidently mistaken machine.

A design company meets the handoff

Aubergine’s current identity makes more sense through that small failure than through a catalogue of technologies. It sells the work of making complicated systems usable, then building the machinery underneath. Its services extend from discovery and research to UX, interface design, web and mobile development, cloud engineering and applied AI. A customer can commission a piece of that work or bring the team through several stages of a product.

The company’s current About page identifies Sarthak Dudhara and Bhakti Dudhara as its founders, describing their return from Silicon Valley and the business’s beginning in 2013. Sarthak is CEO; Bhakti is chief design officer. That pairing expresses the commercial proposition rather neatly: business decisions, design decisions and engineering decisions belong in the same conversation. Each handoff is an opportunity for somebody’s carefully understood problem to become somebody else’s misunderstood specification.

There are plenty of agencies capable of drawing interfaces and plenty capable of writing software. Aubergine’s pitch sits where those markets overlap. A buyer comparing it with a specialist design studio, an engineering contractor or an internal hire should ask who will carry the original user problem into implementation. Keeping those responsibilities together is the company’s proposed advantage; the portfolio supplies the more interesting evidence.

The patient, the vehicle, the spreadsheet

Consider ASENT, built with Deccan International for non-emergency patient transport. Scheduling requires matching appointments, patients, vehicles, staff and equipment. A transport slot is only useful if the right vehicle and assistance can arrive. Aubergine’s case study describes the friction of manual coordination and scattered information. The problem becomes harder every time the day deviates from the plan.

The response was a shared operational view: a Gantt-style timeline linked to a map, with schedules, resources and compliance information visible together. Operators can adjust a delayed journey without rebuilding the whole day. Aubergine reports a 30% reduction in scheduling time and a 20% decline in operating costs. Those are company-reported outcomes. The iF Design listing independently identifies the project, credits Aubergine and records a 2023 launch.

Fitness has a connection problem, too

Cubii, the connected under-desk elliptical business, brought a different version of the same challenge. Its app had to work for people of different ages, abilities and technical confidence. Aubergine describes starting with a UX audit, feedback and interviews, then improving navigation and visual cues. Bluetooth reliability mattered alongside the interface. A beautiful activity screen is of limited comfort when the machine beside your chair refuses to connect.

The work also included community features, challenges and personalized interactions. That breadth explains the customer fit. Aubergine serves businesses whose product experience runs through several layers at once: device, app, data and daily habit. The paying customer is the product company; the person who benefits may simply want to exercise without becoming a part-time troubleshooter.

The broker gets an off switch

For UAE real estate startup Rezio, Aubergine describes building specialized AI agents for property questions, negotiation and scheduling, coordinated through an orchestration layer. The appealing detail is the control panel. Brokers can allow automatic replies, request suggestions for review or step into a conversation themselves. They can choose which responsibilities to delegate. Automation has adjustable boundaries.

“Combining strategy and action with a focus on quality.”Bhakti Dudhara · Co-founder and chief design officer

Read alongside the Insifi episode, that design becomes a practical buying criterion. Ask where the human takes over, what information travels with the handoff and how completed actions are checked. Aubergine’s October 2026 account of agent failures stresses ownership, evaluation, handoffs and verification. It describes an early tendency to blame the model, then a growing recognition that the surrounding system often caused the failure.

Count the completed job

This is a services business, with scope and continuing delivery shaping the engagement. One historical Clutch client interview records $100,000 spent on a collaboration platform. Treat that as one customer’s bill, rather than a menu price. Research, implementation, integration and maintenance can make superficially similar briefs very different purchases.

Aubergine’s own AI budgeting article applies the same discipline to operating costs. Its illustrative insurance workflow totals about $0.27 per completed claim after multiple model calls, retries and infrastructure. A single-call estimate comes to roughly $0.014. These are worksheet assumptions, not an Aubergine quote. The portable idea is to count a finished business task, including the work needed to check it.

The approach needs usable data, people who can verify answers and a workflow that permits intervention. Without those conditions, adding an agent can multiply uncertainty. Aubergine’s culture page describes Learning Labs, mentorship and learning budgets; its October 2026 announcements add Claude Partner Network Certified status and a Flutter consultant listing. The company is broadening its tools. Its most useful published lesson remains wonderfully small: check the number before somebody uses it.