Every Geschäftsführer in the DACH region has the same question right now: are we ready for AI? The board is asking. Competitors are making announcements. Consultants are circling.
The honest answer is that readiness is not a feeling, and it is not a board resolution. It is a set of observable conditions — and most companies do not have them yet. McKinsey's State of AI in 2025 survey found that 88% of organisations now use AI regularly in at least one function, yet only around a third have begun to scale it across the enterprise. Adoption is nearly universal; production-grade value is not. Roughly 6% of respondents qualify as "AI high performers" attributing 5% or more of their EBIT to AI. The gap between near-universal adoption and that thin band of real impact is exactly where readiness lives.
That gap is not a technology problem. It is an operating-model problem, and it is visible early. After our work with Mittelstand companies, we can read the signals — positive and negative — within the first two conversations, before a single line of code or a single tool decision. Here are the five signs that tell us a company is genuinely ready, and the three red flags that reliably predict failure.
The 5 signs of real AI readiness
1. You can name a specific workflow with volume
Not "we want to use AI in operations." Not "our customer service could benefit from AI." But a concrete statement: "Our procurement team processes 800 supplier invoices per week, and 60% follow patterns that could be automated."
The named workflow is the single most important readiness signal. It means someone has done the hard work of moving from aspiration to specification. It means there is a real process with real volume and real people doing it today. This is also where the value actually lands: in the same McKinsey survey, the AI high performers were nearly three times as likely as their peers to have fundamentally redesigned an individual workflow around AI rather than bolting a tool onto an unchanged process — and workflow redesign was among the strongest single contributors to business impact of every factor tested. You cannot redesign a workflow you cannot name.
Companies that cannot name a workflow are not unready — they are pre-ready. They need Discovery, not deployment.
2. You have an executive sponsor with budget authority
AI initiatives that are "owned" by a committee do not ship. AI initiatives sponsored by someone who says "sounds interesting, keep me posted" do not ship either.
What ships: an initiative with a named executive sponsor — Geschäftsführer, Vorstand, or senior Fachbereich leader — who has direct authority over the workflow, the team that runs it, and the budget to fund the first phase without board escalation.
This person does not need to be technical. They need to be operational. They need to own the outcome, not just endorse the idea.
3. The team has allocated capacity — not just enthusiasm
Enthusiasm is not capacity. When we ask "who will be available for 4 hours per week during the build phase to validate outputs, provide feedback, and test the workflow?" the answer matters enormously.
Ready companies have identified specific people — a team lead, a domain expert, a process owner — who will dedicate time. They have discussed this with those people. They have adjusted workloads or backfilled to make it possible.
Unready companies say "the team is excited" and assume availability will sort itself out. It does not.
4. Data is accessible within weeks, not months
You do not need a data lake. You do not need a centralised data platform. You do not need clean, perfectly structured data.
What you need is access. Can the relevant data — the inputs and outputs of the workflow you named in point one — be extracted, sampled, and reviewed within two to three weeks? Not two to three months of data engineering. Not a data governance project. Just access.
In practice, this means: someone knows where the data lives, has credentials to reach it, and can export a representative sample. If that takes a data governance approval cycle, a platform migration, or a vendor negotiation — you are not ready for this workflow. Pick a different one.
5. Your compliance posture is known — even if it is strict
Companies in regulated industries sometimes assume that compliance requirements make them unready. The opposite is true. A company that knows its compliance constraints — DSGVO obligations, EU AI Act classification, sector-specific supervisory requirements — is more ready than one that has never considered them.
The timing now makes this concrete rather than theoretical, though the dates have moved and it pays to read them correctly. Under the EU AI Act, the transparency duties of Article 50 — labelling AI-generated content, disclosing that a user is interacting with a machine — apply from 2 August 2026. The heavier obligations for the high-risk systems listed in Annex III were originally pinned to the same date, but the Commission's Digital Omnibus has deferred stand-alone Annex III high-risk obligations to 2 December 2027 (a change that becomes legally binding only once published in the Official Journal, expected before August 2026). For a DACH Mittelstand company, the practical question is therefore not whether the regulation exists but whether your named workflow touches a high-risk use case at all — and most back-office automations do not. Knowing that answer in advance is the readiness signal. Known constraints can be engineered around. Unknown constraints cause surprises in month four that derail the entire initiative.
The readiest regulated companies we work with have already involved their Data Protection Officer and legal team. Not for a 12-month review, but for a scoping conversation: "If we automate this workflow with AI, what do we need to address?"
The 3 red flags that predict failure
Red flag 1: Decisions are made by committee
When the question "who decides whether this initiative goes forward?" produces a list of five names and a reference to "the steering committee," the initiative is in trouble.
Committees are excellent for governance. They are catastrophic for execution. AI initiatives require fast iteration, rapid decisions about scope changes, and the ability to kill or pivot a workflow within days. A monthly steering committee meeting cannot support that cadence.
This does not mean governance is unnecessary. It means governance should wrap around an initiative that has a single decision-maker — not replace that decision-maker with a group.
Red flag 2: "We need a data strategy first"
This is the most expensive sentence in enterprise AI. It sounds responsible. It sounds prudent. And it is precisely how companies end up in the majority that have adopted AI but never scaled it — a full data-strategy programme runs for a year or more and produces a document that is outdated by the time it is approved, while not one workflow has shipped.
You do not need a data strategy to deploy your first AI workflow. You need data access for one workflow. That is a tactical problem, not a strategic one.
Companies that insist on a data strategy before any AI deployment are solving a maturity problem when they have a readiness problem. The right sequence is: deploy one workflow, learn what data you actually need, then build strategy from experience rather than theory.
Red flag 3: No one can name the workflow
This is the inverse of sign number one, and it is the most common red flag we encounter.
The conversation goes like this: "We want to be an AI-driven company." — "Great. Which workflow will you automate first?" — Long pause. — "We were hoping you could help us identify that."
Identifying the right workflow is valuable work, and it is exactly what Discovery in The AI Operating System framework is designed to do. But it is pre-readiness work. The company that cannot name a single candidate workflow is not ready to deploy. They are ready to explore.
That distinction matters because deploying without a clear workflow target burns budget and credibility. Exploring with a structured framework produces the workflow definition that makes deployment possible.
How to use these signals
These eight signals — five positive, three negative — are not a scoring system. You do not need all five positive signs to start. But you need at least three of the five, and you need zero of the three red flags.
The most common profile we see in ready companies: they have a named workflow (sign 1), an executive sponsor (sign 2), and accessible data (sign 4). Team capacity and compliance posture are being worked on but not fully resolved. That is enough to begin — those gaps close during the build phase.
The most common profile in unready companies: they have executive enthusiasm but committee decision-making (red flag 1) and no named workflow (red flag 3). That combination predicts a six-month planning phase that produces a strategy document, not a production deployment — and lands the company squarely in the majority that have adopted AI somewhere but scaled it nowhere.
From signals to action
If you recognise three or more positive signs in your organisation, the next step is not more assessment. It is scoping: define the workflow, budget the initiative, assign the sponsor, and start.
Our Diagnostic is designed for exactly this moment — a structured evaluation that takes your named workflow and maps it against the dimensions of operational readiness from The AI Operating System. The output is not a maturity score. It is a deployment plan.
If you recognise the red flags instead, the next step is Discovery, not deployment. And that distinction will save you a year and a six-figure budget spent producing a strategy document no one ships.
A Fit Call pressure-tests your readiness against these eight signals in thirty minutes — so you start the workflow you can actually scale, not the pilot that joins the two-thirds that never do.
References: McKinsey & Company, "The State of AI in 2025: Agents, innovation, and transformation," November 2025, https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-state-of-ai; European Commission, "AI Act implementation timeline," https://ai-act-service-desk.ec.europa.eu/en/ai-act/timeline/timeline-implementation-eu-ai-act; EU AI Act, "Article 50: Transparency Obligations," https://artificialintelligenceact.eu/article/50/; Gibson Dunn, "EU AI Act Omnibus Agreement: Postponed High-Risk Deadlines and Other Key Changes," 2025, https://www.gibsondunn.com/eu-ai-act-omnibus-agreement-postponed-high-risk-deadlines-and-other-key-changes/.