How We Turned Our Customer Research into a Concept We Actually Believed In

There was one conversation I couldn’t stop thinking about.

A restaurant operations manager — five outlets, a team of forty — showed us a binder. Laminated pages. Handwritten tick marks. He flipped through it the way someone flips through something they’ve given up on. “The staff don’t follow it,” he said, not with frustration but with the flat certainty of someone who had accepted this as permanent. “They say they do. The binder says they did. But I have no way to know.”

That binder wasn’t a system. It was a record of things nobody could verify.

We heard some version of that conversation across nearly every restaurant we spoke to in Dubai and the other emirates. Not just once. It came up in the first ten minutes, every time. The tools weren’t working. The paper wasn’t working. And nobody was sure if anything was actually happening the way it was supposed to happen.

We had data. What we didn’t have was a concept.

After those early conversations, Ahmad and I had a clear picture of the pain. Restaurants were losing money and consistency because the gap between what managers expected and what staff actually did was invisible. No evidence. No accountability. No way to coach anyone based on real information.

We had data. What we didn’t have was a concept.

We also had our own backgrounds to draw on. Ahmad had spent 11 years running restaurant operations, including building Mr. Briskets from scratch. I had spent 20 years in web and digital. We understood the problem from both ends.

But our first version of the concept was too broad. It was essentially: “AI for restaurant operations.” That covered everything and therefore described nothing. We talked about SOP management, staff training, compliance tracking, predictive analytics, conversational AI — all at once. When we tried to explain it to potential customers, we saw their eyes glaze over around the third feature. When we tried to explain it to ourselves, we couldn’t agree on which part to build first.

The data was telling us something specific. We were answering it with something general.

What changed was a question Ahmad asked after one of our interviews.

We had just spoken to a food safety inspector — someone whose entire job is to determine, on arrival, whether a restaurant has actually done what it claims. He told us he could tell within three minutes whether a checklist had been genuinely completed or just ticked off after the fact. He looked for photo evidence. He looked for timestamps. He looked for the kind of specificity that’s impossible to fabricate.

Ahmad said: “That’s what managers need. Not a report. A record they can trust.”

That sentence reframed everything. We weren’t building a productivity tool. We weren’t building an analytics platform. We were building an evidence layer! A system that made the invisible visible. Every step a staff member completed would be timestamped, attributed, and backed by a photograph where it mattered. A manager sitting in a second outlet could see what was happening at the first one in real time. Not a summary. Proof.

The concept suddenly had a shape: turn any SOP (standard operating procedure) into a digital checklist that staff execute via QR code, with mandatory photo evidence on critical steps and live visibility for managers across every outlet.

That was it. That was the thing we believed in.

What the concept looked like once it crystallized was almost embarrassingly specific.

We weren’t building for all of hospitality. We were building for operations managers in the Gulf who run multiple outlets and have already tried paper-based systems, WhatsApp groups, and generic checklist apps — and found all of them useless because none of them produced verifiable evidence.

The “why this, for these people, in this specific way” became “because in a market where regulatory compliance is intensifying, where staff turnover is high, and where managers can’t be everywhere at once, the single most valuable thing you can give an operations team is a system they can trust. Not a system that tells them what should happen. A system that shows them what did happen.”

We built Exce around that. The AI handles the SOP digitization. Paste a document, get a structured checklist in seconds. The staff execution is deliberately friction-free: scan a QR, complete the task, take the photo. The manager console gives you the evidence feed, the exceptions, the dashboard. Everything connected. Nothing on faith.

If you’re sitting with customer research right now and can’t find the concept in it, here’s what I’d do.

Read back through your notes and find the constraint that kept appearing. Not the pain; the constraint. The thing your customers have tried and abandoned. The thing they’ve accepted as permanent.

Build for that constraint. Don’t build for the whole problem. The whole problem is too big to trust, and your customers won’t believe you can solve it until you’ve demonstrably solved the specific thing they’ve stopped hoping anyone will fix.

That’s where the concept lives. Not in the insight. In the thing the insight made obvious that you kept trying to make bigger.

Shiraz Karar is co-founder of Exce, an AI-powered app that reimagines how restaurants and multi-store units operate. Shiraz and Ahmed Bin Khaled are Cohort 2 alumni of the FRWRDx IDEA Program.

If this resonates, rolling applications for the FRWRDx IDEA Program are open. 14 weeks, 7 milestones, AED 3,000 — and you keep your company.

Apply