Project Charter & Plan
Hotel operators run on shared intelligence: STR for the market, HOTSTATS for the P&L, revenue management for pricing. Restaurant operators fly blind. There is no equivalent in F&B, in Dubai or anywhere in the region.
The US has one reference point, Black Box Intelligence, and it took years of grind to accumulate the data. The UAE market is younger and more disparate, which is exactly why the door is still open. Whoever does the hard work of signing the first operators owns the category.
Guido validated the thesis without hesitation and pressure-tested it in the right places: is anyone already here, why has no one built it, and is there enough comparable density for a benchmark to mean something. The answers held. His one caution is that this is a long, high-touch sell, covered under Risks below.
Where it came from. The idea started with Sudhin Siva, who saw the gap and shaped the give-to-get commercial model. Anindo Banerjee and Daryanand Shetty then blessed the architecture and set the Microsoft Fabric direction. Guido de Wilde validated the thesis, and Harry Johnson sharpened the go-to-market. Their input runs through the sections below.
FoodSight gives UAE restaurant operators the clear view of their market they have never had, turning their own P&L and POS data into dashboards, reporting and a benchmark they can act on today.
To become the definitive intelligence layer for food and beverage in the region, the STR of restaurants, so no operator makes a major decision without it.
Operators will not hand over data for a benchmark on day one. So we do not ask them to. We lead with a dashboard that gives them insight on their own data they have never had, then accumulate toward the benchmark. Every major benchmarking player did the same, entering as a POS or aggregator first.
A strong, standardised template with real presence: infographics, league tables, an executive summary, and budget and prior-period comparisons. Revenue-management models, menu engineering, predictive prompts, and a plain-English "ask your data" layer on Fabric-connected data.
Built on Microsoft Fabric, connected to the operator's own Fabric. An LLM never touches a source database: data flows source, to a curated data lake, to an MCP layer, to the AI and front end. Each operator is fully isolated, raw data is never exposed, and the benchmark surfaces only aggregated figures over a minimum basket.
Operators contribute their data to unlock benchmarks, rather than buy access without contributing. The dashboard is the reason to contribute in the first place.
The thesis is concentration. Land roughly six large operators and the long tail of 1,000-plus outlets follows, because the benchmark only becomes useful once the anchors are in. Dubai first, with micro-market granularity such as JLT versus Marina, then Abu Dhabi and Sharjah.
Illustrative concentration, not audited outlet counts. Traditional Pareto: operators ranked left to right, cumulative line on the right axis. Seven houses carry about 80% of the estate, then a long tail of many small operators makes up the rest.
A small group of senior industry figures is guiding FoodSight in an informal capacity, sharpening the concept, the go-to-market and the introductions. Their involvement is advisory, not a formal board.
Originator. Saw the gap first and shaped the give-to-get commercial model the whole plan rests on. Data and strategy background, and the warm Shamal channel.
Hospitality leadership. Former Marriott Middle East COO. Validates the concept and opens doors to F&B operators.
Value and strategy. Decades in hospitality value creation. Sharpening the go-to-market and connecting the operator and investor network.
Operator lens. Ex-Starwood F&B, now consulting. A grounded read on the concept and the market.
Revenue and valuation. Asset-valuation and revenue specialist, Nitin's RevAI co-founder. A commercial and economics sounding board.
| Risk | Mitigation |
|---|---|
| Operators reluctant to share data | Lead with the dashboard value on their own data. Secure Microsoft architecture, raw data never exposed, give-to-get contributory model. |
| Long, high-touch sales cycle Guido's main caution | Qualify genuine interest in the room and set the next meeting date on the spot. Run a tight, sequenced cohort push rather than a broad spray. |
| "We will build it ourselves" operator IT teams | Position as the benchmark custodian and partner, not a report-builder. Anyone can query their own data; only we hold the cross-operator benchmark. Differentiate on think-with-AI: natural-language query and predictive insight an in-house BI team does not deliver. |
| Thin benchmark validity | Ensure enough comparable concepts per cohort before publishing. Release only aggregated numbers over a minimum basket. Keep the benchmark format stable over time, Gartner-style, so credibility compounds. |
| Credibility from a standing start | Secure a high-profile anchor early so the pitch can borrow its trust, per Guido's advice. |
| Confidentiality in the demo | Keep the line clear: first-party reports show only the operator's own data; the benchmark shows only aggregated figures. Never share commercially sensitive cross-operator detail. Each operator is isolated inside Microsoft. |
| Third-party data sources | Treat data sourcing as a commercial decision, not just a technical one. For any aggregator or provider, ask who owns it, is it free, how is it qualified, and what it costs us in commercial position. |
| Pricing pressure | Deferred by design. Prove concept and demand first, then package. |
| # | Action | Owner | By when |
|---|---|---|---|
| 1 | Trigger the Max introduction. WhatsApp Guido the reminder he asked for, to set the session with his Head of F&B, Max, who can shape the target-operator list. | Nitin | By Thu 14 Aug (session in Sep, when Guido visits) |
| 2 | Bring Shamal in as first data partner. Connect FoodSight to Shamal's Fabric without hosting their data, and use an industry dataset for day-one benchmarking. | Nitin · Ritta | Wed 12 Aug |
| 3 | Take the advisory read from Stefan Breg. Honest read on the concept, design points to sharpen, and a first set of introductions and targets. | Nitin · Ritta | Wed 12 Aug |
| 4 | Finish v1 of the platform. Lift the dashboard to Harry's bar, add the revenue-management models, lock the FoodSight identity, polish ask-your-data and predictive views. | Ritta | By 31 Aug (target) |
| 5 | Build and sequence the founding-cohort list. Firm up the six anchor targets, warm-first, and work them one by one toward a signed data contribution. | Nitin · Ritta | Draft by 24 Aug |
| 6 | Secure a credibility anchor. Land one high-profile operator willing to be named as backing the concept, so every later pitch carries borrowed trust. | Nitin | By mid-Sep (target) |
| 7 | Bring Kevin (Bangkok) in. Once v1 is where Nitin is happy, Harry shares the stack with Kevin for his commercial read and a joint meeting. Bridge to phase-two Thailand. | Harry · Nitin | Post-v1, wk of 1 Sep |