A business that cannot see itself cannot correct itself.
The Data Canvas organises what you measure into the kinds of intelligence a business actually runs on — and, since the AI era arrived, there are four rather than three.
Four kinds of intelligence
Artificial Intelligence. The layer of agents and models the business runs on — and, just as importantly, the context those models are grounded in. The shift most organisations are now making is to treat AI as a portfolio rather than a single bet: a primary partner that carries the daily work and holds the deepest integration with your context, and secondaries admitted deliberately for a specific strength or for redundancy. Keep one primary, so context compounds in a single place. An AI is only ever as good as what it knows about you, which is why the qualitative knowledge of your business matters more in this era, not less.
Performance Intelligence. Are the goals being met? This is the layer that tells you whether the plan is working, by role and by outcome.
Business Intelligence. Is the business healthy? Revenue, margin, cash, retention — the numbers that describe the enterprise rather than any one team.
Operational Intelligence. Is the work flowing? Cycle times, throughput, quality, the operational signals that move before the financial ones do.
Choose few, and choose early
The discipline is not collecting more. It is choosing the few numbers that tell you the truth early — while there is still time to act — and being honest about which numbers you watch because they are comfortable rather than because they are leading.
Which of your current metrics would tell you about a problem this month rather than next quarter? Those are the ones worth the dashboard.
A worked example
Every worked example in the canon uses the same company — Burke & Wills EV, a hypothetical Australian off-road EV maker — so the picture builds across frameworks instead of restarting each time.
# Burke & Wills EV — Data Canvas (MQ5)
Our source of truth: every outcome metric we chase in 2025 — every dollar, every percentage, every unit — has exactly one system that owns it. If we know where the data lives, we can connect to it, track it, and always know where we stand against where we said we'd be. This is the map.
Overview
We're launching the Wills model into the Australian adventure EV market this year: 500 units delivered, $50M revenue, +65 NPS, zero critical quality issues. You can't run a launch that tight on gut feel. So we drew the line — for each number on the plan, which system is the boss? The answer is below, and the rule is simple: if Power BI and SAP disagree on units delivered, SAP wins, because SAP is where a unit is born. Waymaker sits over the top as the performance hub, holding the goals and plans and pulling the live numbers in so the ELT looks at one scorecard, not nine browser tabs.
The Three Types of Intelligence
1. Artificial Intelligence (AI) — agents and models, run as a primary + secondary portfolio
We run a deliberate AI portfolio, not a single bet — the same way we'd keep a primary and a secondary bank.
| Layer | Our choice | Why |
|---|---|---|
| Primary AI partner | Anthropic Claude | Everyday reasoning, drafting, support replies, and the agents embedded across the business — plus the deepest integration with our own context |
| Secondary AI | xAI Grok | Specialist tasks, real-time/social signal, and redundancy if the primary is unavailable |
| Grounding context | Waymaker — Commander Documents, Notes & goals, exposed via MCP | What makes any model smart about us — our plan, playbooks, and engineering specs, kept in one context layer rather than scattered across wikis |
The rule: one primary so context compounds in one place; secondaries earn their slot on a specific strength, not by default. When the Series B room asks how we scale output without scaling headcount linearly, this portfolio is a large part of the answer.
2. Performance Intelligence (PI) — are our people and teams hitting their targets?
| Metric | Year-1 target | Source of truth | Owner |
|---|---|---|---|
| Units delivered (Wills) | 500 | SAP | David Chen (Head of Production) |
| Test-drive → purchase conversion | 35% | HubSpot | James Thornton (CCO) |
| Qualified leads | 2,500 | HubSpot | Lisa Park (Marketing Director) |
| Service partners signed | 12 | HubSpot | James Thornton (CCO) |
| Sponsorships / partnerships | 15 | HubSpot | Lisa Park |
| Goal & waypoint progress (1/12th rule) | On-track | Waymaker | Sarah Mitchell (CEO) |
3. Business Intelligence (BI) — is the whole company healthy?
| Metric | Year-1 target | Source of truth | Owner |
|---|---|---|---|
| Revenue | $50M | Xero (+ Fathom analytics) | CFO |
| Average selling price | $89,000 / unit | Xero | CFO |
| Gross margin | 22% | Xero + Fathom | CFO |
| EBITDA | −$5M | Xero + Fathom | CFO |
| Customer NPS (at scale) | +65 | Delighted | Emma Richardson (Head of CX) |
| Brand awareness (target segment) | 45% (from 5%) | Brand Track | Lisa Park |
| Employee engagement | 85% | Culture Amp | Marcus Webb (People & Culture Director) |
| Exec scorecard (consolidated) | All-green | Power BI | Sarah Mitchell |
The context that grounds our AI
The portfolio above is only as good as what it knows about us. This is the qualitative knowledge — the how, not the how many — and we keep all of it in Waymaker, not scattered across separate wikis. One context layer serves both our people and our AI: Waymaker exposes it via MCP so Claude and Grok answer as Burke & Wills, not as a generic assistant.
| Knowledge | Where it lives | Owner |
|---|---|---|
| Annual plan, goals, role canvases, meeting rhythm | Waymaker | Sarah Mitchell |
| HR policies, brand playbook, launch playbooks | Waymaker Documents | Marcus Webb / Lisa Park |
| Engineering specs, build standards, thermal/IP65 test records | Waymaker Documents | David Chen |
| "Would you take it to the outback?" quality standard | Waymaker Documents | David Chen |
| Service-partner onboarding & dealer playbook | Waymaker Documents | James Thornton |
The Data Canvas — systems by operational area
We map every tool we run from the inside out: employee experience, then the operational core, then the two intelligence layers, then customer experience. Each system is tagged with the intelligence it feeds.
Employee Experience
| System | Purpose | Intelligence |
|---|---|---|
| Culture Amp | Engagement, key-talent retention (90%), adventure-program completion | PI / BI |
| Waymaker Documents | Policies, onboarding, the four values, "Join the expedition" EVP | AI context |
| Waymaker | Role canvases, 1:1 cadence, fortnightly manager 1:1s | PI / AI context |
We're hiring 45 people this year and we want 85% engagement while we do it — Culture Amp is the only system that gets to say whether we hit that.
Operational Systems (the core)
| System | Purpose | Intelligence |
|---|---|---|
| SAP | Production volume, quality (zero critical issues across 500 deliveries), line efficiency | PI / BI |
| Xero | Finance ledger — revenue, ASP, margin, EBITDA, runway against the 24-month clock | BI |
| Fathom | Financial analytics layer on top of Xero | BI |
| Waymaker Documents | Engineering source of truth (−10°C to +55°C, IP65, 800km+ range, 3,500kg towing) | AI context |
SAP owns the unit. A Wills isn't "delivered" until SAP says so — that's how we'll know we hit 500, and it's how we'll prove zero critical issues to the Series B room.
Performance Intelligence systems
| System | Purpose | Intelligence |
|---|---|---|
| Waymaker | Goal tracking, waypoints, ELT scorecard, plan-to-actuals | PI |
| HubSpot | Lead pipeline, conversion, partner & sponsorship tracking | PI |
Business Intelligence systems
| System | Purpose | Intelligence |
|---|---|---|
| Power BI | The exec dashboard — single pane for the Monday ELT sync | BI |
| Fathom | Revenue/margin/EBITDA trend analytics | BI |
| Brand Track | Unaided awareness in the 35–55 adventure segment | BI |
Power BI is the picture; it doesn't own a single number. Every tile on it traces back to SAP, Xero, HubSpot, Delighted or Brand Track. If you want to change a number, you change it at the source, not on the dashboard.
Customer Experience
| System | Purpose | Intelligence |
|---|---|---|
| HubSpot | CRM — every lead from first touch to the 35% who convert | PI |
| Delighted | NPS source of truth (+65 target at scale; beta ran +85) | BI |
| Zendesk | Support tickets across the 12 regional service partners | PI |
| Brand Track | Brand health and awareness for "Adventure Alex" | BI |
Our customer is Adventure Alex — 42, $180,000 household income, six to ten outback trips a year. HubSpot tracks how we win him, Delighted tracks whether he'd recommend us, Zendesk tracks whether the network keeps him moving. Three systems, one customer, no overlap.
Data Integration Architecture — Waymaker as the hub
We run a deliberate hub-and-spoke model. The specialist systems stay specialist — SAP makes vehicles, Xero keeps the books, HubSpot runs the funnel — and Waymaker is the performance hub that holds our goals and plans and pulls the live actuals in beside the targets. Power BI is the exec read-out for the Monday sync; Waymaker is where the goal, the owner, and the number live together.
```
SAP ──units, quality──┐
Xero + Fathom ──$, margin, EBITDA──┐
HubSpot ──leads, conversion, partners──┤
Delighted ──NPS──┤ ┌──> WAYMAKER (performance hub: goals, plans, scorecard)
Brand Track ──awareness──┤ │
Culture Amp ──engagement──┘ └──> Power BI (exec dashboard, Monday ELT sync)
Waymaker Documents & Notes ──the context layer that grounds Claude + Grok (via MCP)──
```
The rules of the road:
- One owner per metric. No number has two sources of truth. SAP owns units; Xero owns dollars; Delighted owns NPS. Disputes resolve to the owner system, always.
- Source systems feed up; the hub never feeds down. We never write a corrected figure back into the dashboard. Fix it at SAP/Xero/HubSpot and let it flow.
- Targets and actuals sit side by side. Every goal in Waymaker carries its target (500 units, $50M, +65 NPS, 45% awareness) and its live actual from the source system, so the gap is always visible.
- 1/12th rule on the timeline. Annual targets break into monthly waypoints in Waymaker, so by month's end we know if 42 units delivered is on pace for 500.
- No shadow scorecards. If a team wants to track something new, it gets a source of truth and a home in this canvas first — we don't run numbers out of a private spreadsheet.
How the canvas feeds the operating rhythm
| Meeting | Cadence | Looks at | Owner |
|---|---|---|---|
| ELT weekly sync | Mondays 9:00–10:30 | Power BI scorecard + Waymaker goal progress | Sarah Mitchell |
| Daily production standup | 7:00, 15 min | SAP line + quality | Production Supervisor |
| Monthly commercial review | First Friday, 90 min | HubSpot pipeline, Xero/Fathom, Delighted | James Thornton |
The Monday sync runs Wins / Scorecard / Goal progress / Roadblocks — and every one of those four agenda items is backed by a system on this canvas. Nobody walks in with a number they can't trace. That's the whole point: when the Series B room asks "how do you know," we point at the source, not at a slide.