You already know something needs to change. This is the step that turns that into a sequence: what to do first, what it should cost, how to explain it to the people who sign the cheque — and where AI agents fit in it.
We take your free diagnostic results and a short structured intake, then turn them into a step-by-step transformation plan: what to fix first, what to fix next, and roughly what each step should cost. You get a document built for a boardroom, not a technical audience — including where the AI agents will pay off first, and when.
| Deliverable | What it does for you |
|---|---|
| Reference architecture | The target data and technology architecture your deployment should follow — mapped to your actual systems, not a generic diagram. |
| The written roadmap | A prioritized, sequenced plan — what to do, in what order, and why — with a rough cost band for each step, in language a board can approve without a translator. |
| Vendor-neutral tool & integrator shortlist | Categories of technology to evaluate and potential integration partners to get quotes from — never a single favorite, and never anyone we're paid to recommend. |
| Quick-win use cases | A short list of early wins to start proving value with stakeholders while the foundation work is underway — not everything has to wait for phase one to finish. |
| A 60-minute readout | A live working session to walk the plan, answer questions, and pressure-test it against your real constraints before anyone signs off on it. |
| A board-ready summary | A short, forwardable version built for the people who weren't in the room — the ask, the sequence, and the payoff, on one page. |
What you do, what we do, at each step — no discovery-phase mystery.
An excerpt from a fictional plant, so you can see the shape of it before you commission one.
| Initiative | Why now | Est. cost band | Depends on |
|---|---|---|---|
| 1 · Connect & collect on Line 4 | Nothing downstream is trustworthy until the data leaves the floor cleanly. | $40–70k | — |
| 2 · Unified namespace + trusted model | Kills the SAP/Excel/MES silos that make every OEE number arguable. | $60–110k | Phase 1 |
| 3 · Real-time plant view | One pane the morning meeting can act on, instead of five dashboards. | $30–50k | Phase 2 |
Sequencing note. AI pilots are deliberately held until phase 4 — after the data is contextualized and trusted. Piloting sooner produces confident, wrong answers and rework.
"Fund phase 1 now: it's the cheapest step, unblocks every later one, and turns numbers nobody trusts into numbers the board can. Full spend is banded and staged — no blank cheque." — the one-line board summary
Instead of a pile of vendor pitches and internal opinions about what to fix next, you have one document that ranks the work, prices it in bands your finance team can sanity-check, and gives you the exact language to bring to the board or the executive team. You use it to greenlight step one, hold vendors to a sequence instead of their sales pitch, and measure progress against a plan you already approved.
A credible, fundable plan needs real inputs — this covers considerably more ground than the free diagnostic. Illustrative examples below; the real one is tailored to your diagnostic results.
| Category | Example question |
|---|---|
| Why now | What's driving this now — and what's already been tried that didn't stick? |
| How work actually happens | Walk us through a typical shift or order, start to finish — where do people wait, guess, or re-key information? |
| What's already automated | How much of your production data arrives on its own vs. gets typed in after the fact? How far do people actually trust the numbers? |
| Systems already in place | What tracks equipment status on the floor today? What holds your production schedule and inventory? |
| Where it gets expensive | Which process costs you the most when it underperforms, and roughly how much? |
| Who'd need to be involved | Who would champion this internally, and how well do teams currently share information with each other? |
| Network & security basics | How separated are your plant-floor and office networks today, and who owns that decision? |
| What you can commit to | What budget band and timeline would your board realistically approve, and who signs off? |
How you'll answer. Once you commission the plan, open the real intake from your account. This is a real intake, not a quick form - expect it to take a couple of business days to answer thoroughly. Some questions are best answered by someone else on your team; delegate freely, a complete answer beats a fast guess.
If you haven't yet run the free diagnostic, start there — it's free, takes about ten minutes, and tells you whether the roadmap is worth commissioning yet. The roadmap is the step after that — something you can hand to a board, a CFO, or a steering committee and get a yes. It is a one-time purchase on its own: it does not require dxpert Pro, and dxpert Pro does not require it.
No subscription, no retainer, and no dxpert Pro required. You get the architecture, the roadmap, the shortlist, the use cases, the readout, and the summary — then you decide what's next.