AI Readiness Assessment: How to Know If Your Business Is Actually Ready for AI
An AI readiness assessment scores whether you have a workflow worth automating before you pay for models. Here is the seven-question check webvise runs, with a real example.
An AI readiness assessment scores three things before you spend money: whether you have a workflow worth automating, whether the data it needs is reachable, and whether someone will own the result. Most companies that feel behind on AI already have the budget and the tools. What they lack is one named workflow worth the build.
You signed up for three AI tools last quarter. Two of them sit unused because nobody mapped them to an actual job.
That gap is normal. The tools shipped faster than anyone could decide where they fit. This article gives you the seven-question readiness check webvise runs before any AI build, the table we score against, and a real example of what ready looked like for a 25-year-old construction firm.
Readiness is about a workflow, not a shopping list. Name the task that eats hours before you pick a model.
The data has to be reachable. You rarely need a full RAG pipeline before you can ship something useful.
A good assessment can end with "don't build yet." That answer still saves months of wasted budget.
webvise runs a paid readiness assessment from €2,500 over 2 to 4 weeks, ending in a roadmap you can act on with us or without us.
What an AI readiness assessment actually checks
Most AI projects stall because they start at the model and look for a use case afterward. A readiness assessment runs the other direction. It starts at the work, finds the few tasks where AI pays back, and only then asks which tools fit.
Three checks carry most of the weight: a workflow you can name and measure, data a machine can reach, and a person who will use the result every day. When one of those is missing, the technology question does not matter yet.
The seven questions below come from production, not a slide deck. webvise's founder works as an AI engineer at luca, a Berlin fintech, shipping LLM systems daily. Each question on the list maps to a way one of those projects could have broken.
If you want that scoring done for you, webvise's AI consulting service delivers a readiness assessment as a fixed engagement, with a written report and a prioritized list of what to build first.
The seven-question readiness check
Score each question for the workflow you have in mind. A workflow that lands on the right column twice is rarely ready. One that sits in the left column across all seven is worth a build conversation this quarter.
| Question | Ready when | Not ready when |
|---|---|---|
| Is there a named workflow? | You can describe one task and the hours it eats each week | AI is a goal with no target task |
| Can you measure it? | You know the volume and what an hour of it costs | You only sense that it feels slow |
| Is the data reachable? | The inputs live where an API or export can read them | The knowledge sits in people's heads and scattered PDFs |
| Is the process stable? | The steps were the same last quarter | The workflow changes every few weeks |
| What happens when it is wrong? | A human reviews the output before it acts | An error ships straight to a customer with no check |
| Who owns the result? | One named person will use it daily | Nobody is accountable for adoption |
| Can you see payback? | Hours saved times cost beats the build cost | ROI is a feeling rather than a number |
The payback question stops more projects than any other. A task that runs twice a month rarely earns back the cost of building and maintaining an AI system around it. The math works once a workflow happens daily, touches many people, or blocks revenue while it waits. Our AI automation ROI calculation walks through the numbers with a worked example.
What ready looked like for a Brandenburg construction firm
A construction company in Brandenburg, founded in 1999, came to webvise for a website rebuild. The readiness questions reshaped the brief. Their highest-value workflow was answering the same project and service questions over and over, for an international workforce that spoke several languages.
That workflow passed the check. The data was reachable because their project portfolio and service descriptions were already structured content. The task was stable, it happened constantly, and a wrong answer was low risk because a person handled anything the bot could not.
We shipped a Next.js platform with a chatbot built on Google Gemini 2.5 Flash for instant visitor queries, in 8 languages, plus an email-to-issue pipeline that turned a forwarded email into a content update. Delivery took 3 weeks. The site scores 95 on Lighthouse performance and loads in under 1.5 seconds. The assessment is what kept the AI scoped to one job that paid back, instead of a generic feature nobody asked for.
The most common reason the answer is not yet
Data reachability fails more readiness checks than anything technical. Teams assume the AI cannot help because the knowledge is messy. The real blocker is that the knowledge has no door a machine can open.
This trips up a lot of buyers who think step one is an expensive vector database. Often a simple export, a shared folder, or a structured table is enough to start. We cover when you can skip the heavy retrieval stack in building a business knowledge base without RAG.
Process stability is the second quiet blocker. If a workflow changes every few weeks, any automation you build is out of date by the time it ships. Sand the process down to a stable shape first, then automate it. A messy process automated is just a faster mess.
What a readiness assessment costs, and what you get
webvise runs the assessment as a fixed engagement from €2,500 over 2 to 4 weeks. That covers workshops with your team, the scoring you saw above applied to your real workflows, and a written roadmap. The cheapest outcome is the one that tells you not to build, because it saves the months a wrong bet would have cost.
| You leave with | What it answers |
|---|---|
| Readiness assessment report | Which of your workflows pass the seven checks |
| Prioritized use case catalog | What to build first, second, and never |
| Architecture recommendation | Which models and tools fit, and the reason |
| Implementation roadmap | Milestones, cost ranges, and who does each step |
The roadmap is built to be acted on by anyone. Many clients move straight from the assessment into webvise's AI automation service to build the top use case. Others hand the roadmap to their own team. Both are fine, because the value is the decision, not the lock-in.
Run a lightweight version yourself this week
You do not need a consultant to start. Pick the workflow that annoys your team most and run it through these four steps before you book anything.
Write the task in one sentence. If you cannot, it is too vague to automate yet.
Put a number on it. Hours per week times the loaded cost of the person doing it.
Find the data door. Name where the inputs live and how a machine would read them today.
Name the owner. Write down the one person who will use the result, or stop here.
If three of the four feel solid, you have a candidate worth a real assessment. If you are still deciding whether to bring in help at all, our guide on AI consulting for small business covers when a packaged tool is enough and when an engineer should get involved.
webvise scores AI readiness the same way we ship production systems: start at the work, prove the payback, then build. If you want the seven questions run against your own workflows, tell us what eats your week at our contact form.
Webvise practices are aligned with ISO 27001 and ISO 42001 standards.