A sale. A purchase order. A stock move. A late truck. A payroll run. We turn each one into money, in the same model, in real time. F&B, Retail, Manufacturing, Logistics. Outlets, stores, plants, depots, all on one ledger your team and your agents can both read.
Outlets, channels, recipes, aggregators. Ingredient-level COGS and AOV by daypart.
Stores, DTC, marketplaces, returns. SKU-level margin, basket mix, and shrink tracked per door.
Plants, BOMs, shifts, suppliers. Yield, OEE, raw-material pass-through and standard-cost variance.
Depots, fleets, lanes, 3PLs. Cost-per-km, on-time delivery, fuel exposure, and utilization by asset.
You have great tools, POS, ERP, WMS, TMS, HRMS, aggregators, marketing. The picture of the business lives in someone's head, taped together with WhatsApp and overnight Excel. That is where the money leaks.
Aggregator recon, discount abuse, over-portioning, and stock shrinkage, individually small, collectively a margin killer.
Numbers land in the board deck on D+10, too late to act, too old to trust. Founders chase CFOs. CFOs chase analysts. Everyone waits.
Ops says one revenue number, finance says another, and the investor deck says a third. Cohort, SSSG, AOV, defined differently in every file.
A sale. A purchase order. A stock move. A late truck. A payroll run. A customer return. We turn each one into money, in the same model, at the same time, in real time. The ledger becomes the grammar of your data, because every team already speaks it.
It stores rows. You still have to stitch them into a story. The CFO is still waiting on D+10.
It knows facts. It does not know what each fact is worth. An agent can read it and still be wrong.
Every node has a value. Every edge moves value. Ask any question, get a number, with a row that proves it.
Vector search. It finds text written by people. People hedge, summarise, and forget. Your agent guesses, and the answer changes every time you ask.
The Economic Graph. Same question, same answer, every time, with a row that proves it. Your agent stops guessing and starts knowing.
A real 12-week plan in Pilot. Actuals (W-1 to W-3) are locked. Grab any forecast point on the poultry price line, drag it up or down, and every KPI, every table row, every chart refolds in place, when you publish, Pilot writes the plan back to the ledger.
| Actuals | Forecast | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Cost of Goods | W-1 | W-2 | W-3 | W-4 | W-5 | W-6 | W-7 | W-8 | W-9 | W-10 | W-11 | W-12 |
| Price per KG | 200 | 200 | 200 | 200 | 200 | 200 | 200 | 200 | 200 | 200 | 200 | 200 |
| Change | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | |||
| Weekly Usage | 64.4 | 63.1 | 59.3 | 56.9 | 64.2 | 63.1 | 59.7 | 56.9 | 61.6 | 67.7 | 58.8 | 55.9 |
Illustrative F&B plan. Your chart of accounts, your drivers, your weekly feed plug in during onboarding.
Drafts, branches, diffs, reviews, publishes, reverts, blame, treated as first-class primitives, but for reports, forecasts, and operating plans, not code. One plan is a branch. Saving creates a commit. Comparing two drafts is a diff. Publishing is a merge. Every cell carries its own blame.
Spin up a draft off the approved plan, price scenario, new outlet ramp, promo calendar, hiring freeze, without touching anyone else's numbers.
Save any state as a named version, v1 +5.5 USD/kg poultry, v2 fuel shock, v3 aggressive promo. Each commit is an immutable snapshot the board can reference.
Compare two drafts side-by-side on revenue, margin, EBITDA, and cash, with the drivers that caused each delta surfaced, not buried in a 40-tab Excel.
Inline comments on any cell. CFO or Ops head approves the plan, the review chain is captured on the commit, same discipline as a pull request.
Promote an approved draft to the company of record. Dashboards, the weekly MIS, and the board pack all rebase on the new plan instantly.
Roll back to any prior version in one click. Hover any number to see who moved it, when, why, and which version introduced the change.
Same discipline a great engineering team puts into code, now applied to the weekly MIS, the board pack, and the three-year plan.
Most teams don't see indicative numbers until day 7 or 10 of the following month. By then the month is done, and so are the decisions. DataHQ's P&L renders as the business happens, and the driver simulator above is writing straight into this sheet.
| Line item | Baseline | Scenario | Δ |
|---|---|---|---|
| Revenue | 45.00M | 45.00M | 0 |
| Dine-in / DTC | 28.80M | 28.80M | 0 |
| Aggregators / wholesale | 13.50M | 13.50M | 0 |
| Other & catering | 2.70M | 2.70M | 0 |
| COGS | (13.50M) | (13.50M) | 0 |
| Food · poultry, dairy, produce | (2.97M) | (2.97M) | 0 |
| Other food & beverage inputs | (8.73M) | (8.73M) | 0 |
| Packaging | (1.15M) | (1.15M) | 0 |
| Wastage & loss | (0.65M) | (0.65M) | 0 |
| Gross profit | 31.50M | 31.50M | 0 |
| Gross margin | 70.0% | 70.0% | 0 bps |
| Operating expenses | (26.10M) | (26.10M) | 0 |
| Rent & occupancy | (6.30M) | (6.30M) | 0 |
| Labor & benefits | (10.10M) | (10.10M) | 0 |
| Aggregator & channel fees | (4.90M) | (4.90M) | 0 |
| Marketing, utilities, G&A | (4.80M) | (4.80M) | 0 |
| EBITDA | 5.40M | 5.40M | 0 |
| EBITDA margin | 12.0% | 12.0% | 0 bps |
One graph underneath. Three products on top. Fabric builds the graph, every event from your systems is turned into a row, a value, and a link. Spyke watches the graph, reports, dashboards, alerts, and the weekly and monthly MIS your board reads. Pilot reasons over the graph, simulate a plan, model a scenario, ask in plain English, and write the answer back, audited.
System-agnostic, always-on. Every invoice, shift, PO, production run, and POS ticket, captured on our headless event API and reconciled into one model that mirrors how your business actually runs, end to end.
The narrative layer on top of the ledger. Spyke turns Fabric into insightful reports, live dashboards, programmable alerts, and the weekly and monthly MIS your board actually reads, every number traceable back to the source row.
Plain English in, chart and SQL and footnote out, plus a plan you can actually change. Pilot lets you simulate a scenario, model a driver, and ask anything in natural language, and when you are ready, it writes the answer back into the ledger, audited, versioned, owned.
Cashflow is a data problem. Not a forecasting problem, not a spreadsheet problem. The moment you change a driver in the simulator above, every invoice date, every payment term, every advance, and every receivable recalculates, and you see exactly when the cash actually lands, or doesn't.
| Week | Inflow | Poultry payable | Other outflow | Net | Closing balance |
|---|
A generic AI does not know your sites, your SKUs, your shifts, or that "SSSG" in your board deck excludes stores under 13 months old. DataHQ is the Economic Graph, a live model of every sale, PO, stock move, and payroll run, all turned into money. Plug your AI, your BI, your agents into it, and the answers stop drifting.
The plan is a living document, not a monthly archaeology dig. Spin up drafts for a price change, a new outlet, a promo, or a hiring freeze. Compare two side-by-side. See the bottom-line delta in real time. Publish and Pilot writes the approved plan back to the ledger.
DataHQ auto-discovers the real process from your source systems: POs, GRNs, invoices, stock moves, POS tickets. Frequencies, bottlenecks, and rework show up as edges and colors, so you fix the real bleed, not the imagined one.
Actual, Plan, Forecast and Prior Year, same chart, consistent notation, tabular numerals. Your CFO can read it. Your operator can act on it. Your investor can paste it into a board deck.
| Outlet | Net sales | SSSG |
|---|---|---|
| Marina Walk | 84,201 | +9.4% |
| JLT T3 | 71,882 | +7.1% |
| DIFC Gate | 68,104 | +4.2% |
| Al Barsha | 62,430 | −1.8% |
| Mirdif City | 58,920 | +2.6% |
We don't replace your POS, your accounting, or your HRMS, we reconcile them into one model. Your data stays yours. Every surface, dashboards, reports, AI answers, runs off the same semantic core.
Because the operator's question is not the investor's question is not the franchise partner's question, and all three need to agree.
All sites, channels, and entities in one P&L, closed D+3.
Outlet, store, plant, or depot, each with its own P&L, cohorts, and LFL.
Dine-in, delivery, marketplace, wholesale, lane, margin, not just revenue.
Food:Bev, apparel:hardlines, assembly yield, the lever most ignore.
First-visit retention, by site, by channel, mobile-number grade.
Every KPI against plan and forecast, with a CFO-grade one-liner attached.
Board-grade charts, LFL excluding new sites, MoM + YoY, footnoted.
"Why was Friday light?" / "Why did Plant 2 miss OEE?", chart + cause + owner.
ErasiedtoX is our in-house forecasting model, trained on the physical-world signals most forecasts ignore: coffee bean prices, weekend weather, Ramadan, steel and resin indices, diesel and freight rates, apparel FOB costs, an aggregator down-ranking your outlet. It wires those drivers in, and gives you intervals, not a single false-precision line.
We closed April on the 3rd of May. First time in six years. My board meeting now starts with the numbers, not with an apology for them.
Built by people who've run stores, closed books, and shipped data products at scale, and got tired of doing it in 40-tab Excel.
Blends AI, process mining, and non-ergodic planning across F&B, Retail, and Airlines. Races motorcycles; swims open sea.
20+ yrs of product R&D and high-throughput platforms. Metrics-driven; .NET Core, Angular, Azure ML, Edge.
Three decades in tech; builds enterprise SaaS GTM for Data and AI. Drives triple-digit YoY growth through customer-centricity.
Two decades across fintech, payments, and enterprise, Cashfree, Razorpay, Mswipe. Believes finance needs a platform, not more spreadsheets.
30 minutes. Connect one source (POS, ERP, WMS, TMS) and one accounting system. We'll show you the site-wise P&L your CFO has been asking for.