Question 1

What actually is a digital twin? Everyone defines it differently.

Accepted Answer

A digital twin is a digital copy of something real — a plant, a network, a building, a body — that shows what has happened, what is happening, and different versions of what could happen next. Some of the most valuable twins exist before the real thing does: you rehearse the building before you pour it, the network before you deploy it.

What makes it a twin is not the technology. It is that people, systems and AI can finally reason over the same reality together, and that the model can grow in fidelity and twinning rate as your decisions demand it.

That starts with a shared picture: the point where owner, operator, engineer and vendor agree on what exists. Many organisations already have that and don't realise how far along it puts them. It doesn't need to be live — a weekly scan is enough for most maintenance planning. Connect it and it becomes a connected twin. Let it act back on reality, with people in control, and it moves toward autonomous operations.

Each stage pays for itself. Value starts at the shared picture, not at the sensors.

Question 2

How do I know we won't be locked into you?

Accepted Answer

Because we designed the business so you can fire us. No lock-in, stop anytime, and every project ships with training so your team can run it after we leave. The goal is capability in your house, not a permanent retainer.

Technically it's enforced, not promised: open protocols, open formats, everything replaceable. If a component of what we build can't be swapped out for an alternative, we consider it broken.

Here is the honest business logic: consultants who create dependency get one big client and a reputation problem. Consultants who create capability get referrals. We prefer referrals.

Question 3

What is the SMILE methodology in one answer?

Accepted Answer

SMILE inverts how technology projects usually run. Instead of data first — collect everything, build the lake, hope insight emerges — it runs impact first, data last: start from the outcome, work backwards through action, insight and information, and only then decide what data you actually need. Usually it's a fraction of what a data-first project would have hoarded.

Six phases, from emulating reality through to perpetual wisdom, viewed from three perspectives: people, systems, planet. It's published with a DOI (10.5281/zenodo.21268264), there's a free scoring tool on the site, and every WINNIIO engagement runs on it.

Why publish the method openly? Because the methodology only works when your team owns it. A secret methodology is just lock-in wearing a lab coat.

Question 4

What does it actually cost to work with you?

Accepted Answer

Public numbers, on the website: AI agents from €1,500 a month, fractional AI leadership from €3,000 a month, advisory at €500 an hour, keynotes from €2,500. Twin projects are fixed scope — you get a quote in 48 hours, not a discovery phase that bills you for six months.

Why publish prices? Because hidden pricing is a sales tactic, and sales tactics are a tax on trust. You should know before the first call whether we are in your budget.

Startups get a discount. Just ask.

Question 5

Will AI replace our engineers?

Accepted Answer

No. It will replace the parts of their job they hate.

Every hour we automate is time handed back to a human being. Your best engineer currently spends most of her week on status reports, data wrangling and integration plumbing — work an agent does happily at 3 a.m. What agents cannot do is know your plant the way she does. That knowledge is your actual moat, and AI makes it more valuable, not less, because it can finally be applied at scale instead of being trapped in one person's calendar.

The move toward autonomy is a promise: to make us more human, not less. The companies that get this wrong will use AI to cut people. The companies that get it right will use AI to multiply them. Guess which ones keep their best engineers.

Question 6

Does NIS2 apply to us, and what should we actually do about it?

Accepted Answer

If you run energy, transport, water, health, digital infrastructure or manufacturing of critical goods in the EU — assume yes, and assume your OT is in scope, not just your IT.

The good news: NIS2 mostly demands what you should want anyway. Know what exists, how it's connected, who can access it — that inventory discipline is both the first step of compliance and the first step of any twin project. Same work, twice the payoff. A digital twin of your infrastructure is, among other things, a living asset register: the thing every auditor asks for first and almost nobody has.

What we'd avoid: buying compliance as a binder of documents. Paperwork describes your security posture. Architecture is your security posture.

Question 7

Is there a maturity ladder for digital twins? Where would we be on it?

Accepted Answer

Yes, and most organisations are lower on it than their vendors told them.

The ladder in SMILE terms: first emulate reality — get an honest, current representation of what exists. Then connect it — live data flowing in. Then reason over it — simulation, agents, what-if. Then let it act — autonomous operations with people in control. Each rung has to be earned; skipping one is how twin projects die.

Where are you? Run the SMILE score on the site — it walks your situation through the phases and gives you a number, element by element. We did exactly that for a European underground lab: readiness scored element by element, so value arrives in the opening weeks rather than at the end of a 32-month plan.

Question 8

Wireless or wired for building sensors?

Accepted Answer

Both — by layer, not by ideology.

Wired for the backbone: HVAC control, life safety, anything where a failure has physical consequences. That world is conservative for good reasons — it has earned its caution. Wireless for the edges: the sensing layer, where Matter and Thread have finally grown real commercial traction after years of being a spec without products. A mesh of small, battery-efficient devices talking locally, without a single vendor's cloud in the loop — that's now a legitimate option where six years ago it was a promise.

The architecture question matters more than the medium: local control first. The building must be able to run itself without a round trip to anyone's server. Get that right and the wired-versus-wireless debate mostly resolves itself.

Question 9

How worried should we be about our building systems being hacked?

Accepted Answer

Worried enough to act, not enough to panic. Here's the claim: boilers are the new printers. In 2020 the unpatched printer was the easiest way into a corporate network. In 2026 it's the boiler, the chiller, the access panel, and the remote-support tunnel a controls vendor left open two winters ago. Ransomware operators have hit building automation directly — hospitals, campuses, towers — because it's easier to reach than core IT and just as effective at forcing a payout.

The fix starts with inventory, not products: what exists, how it's connected, who can access it. Skip any of those three and the rest is theater. Most building security failures aren't sophisticated — they're an inventory gap that became an entry point.

One attack surface, defended as a whole. Attackers don't respect your org chart.

Question 10

Be honest: is the metaverse dead?

Accepted Answer

The costume party is dead. The infrastructure won.

What died was the pitch: legless avatars in meeting rooms, virtual real estate speculation, a headset-shaped solution hunting for a problem. What survived — and is quietly compounding — is the industrial version: BMW building factories in simulation before building them in steel, network operators rehearsing changes in physics-grounded twins, storm damage assessed from photoreal capture in hours instead of weeks. Same underlying technologies. Opposite direction: instead of escaping reality into a virtual world, you bring reality into the computer so you can make better decisions about it.

We were on stage arguing this — digital twins and the metaverse as tools to improve reality — back when the hype was at full volume. The word was always the least interesting part. Call it the industrial metaverse, the spatial web, holographic societies. The name will keep changing. The capability is permanent.

Ask WINNIIO anything.