What does an AI audit or AI roadmap engagement cost, and what do you get?

Short answer: An AI opportunity audit (also sold as an AI roadmap or readiness assessment) costs about €1,500–€4,000 as a one-day workshop and €6,000–€15,000 as a multi-week analysis in Germany. You get a ranked list of what to automate first. Governance audits against ISO/IEC 42001 are a different, pricier product. Made By Agents' Deep AI Audit is a fixed fee, quoted after a free Discovery Call.

Last reviewed: 2026-10-31. Market prices were read from each provider's own page on that date. They are list prices, net of VAT where stated, in the currency the provider publishes.

Search for "AI audit" and most results are about governance: checking whether AI systems are fair, documented and compliant. Most owners and COOs asking the question mean something else: where does AI pay off in our business, and what should we build first? Know which one you are buying.

If you have no AI systems in production yet, start with an opportunity audit. A governance audit has little to look at until you do.

The pattern: fixed fees, one to three weeks, and a credit against the build are now standard for opportunity audits. Governance work costs more because it involves accredited auditors and repeat surveillance.

The published prices above differ by a factor of ten. These are the reasons:

1. Scope. One workflow, one department or the whole company. Most fixed-price offers state a limit, for example "up to three processes" or "one or two workflows".
2. Depth of the output. A workshop produces a prioritised list. A multi-week audit adds process maps, data checks, build-vs-buy calls and a spec for the first system.
3. Systems and data access. Checking whether your ERP, CRM or DMS can actually be connected takes longer than asking about it, and it's the part that changes decisions.
4. People involved. More process owners, sites and languages mean more sessions.
5. On-site or remote. Travel adds cost. Most providers quote remote work by default.
6. Opportunity or governance. Governance audits involve accredited auditors, documentation reviews and repeat surveillance, so they sit in a different price band.
7. Credit against the build. Many providers credit the fee if you go ahead within 60–90 days, so the effective cost depends on what you do next.

• Price: Made By Agents doesn't publish prices. The audit is a fixed fee, quoted after a free 20-minute Discovery Call once the scope is clear. If you start implementation with us within 90 days, the audit fee is subtracted from that project.
• Duration: most audits take 1–2 weeks after the Discovery Call, followed by one walkthrough session.
• Your time: a few hours with the people and systems that run the process. Walkthroughs, sample files and architecture notes are usually enough. NDAs are standard.
• Who does it: Tobias Wupperfeld, founder, with more than 5 years of AI development experience. He takes every Discovery Call and runs the audit himself. Focus areas: quote and document automation, agents and internal tools, mobile and web applications, and build-vs-buy decisions.
• Remote or on-site: remote by default. On-site is possible if you cover the travel costs.
• AI agents in our own work: we use AI agents in the audit itself, so everything that can be generated (process maps, the analysis write-up, the spec draft) comes together quickly. The time goes into talking to your people and checking the results.
• Already have AI projects underway? We look at what you're doing, tell you what's worth keeping, and suggest what to do next.

1. Analysis of the status quo. How you handle quotes, documents, knowledge, reporting or internal tools, including the steps that slow them down.
2. What to build first. Three to five systems, in the right order: impact, effort and internal dependencies.
3. A spec for the first system. Who it's for, what it connects to, and what production-ready looks like.
4. Build vs buy. Where an off-the-shelf tool is enough, where custom software pays off, and where you should wait.
5. A walkthrough with the process owners. You decide what happens next: build in-house, book a consulting session for your team, or hire us for the first system.

"Made By Agents has been an invaluable partner in our digital transformation. They helped us identify and implement the exact AI tools needed to automate processes that were previously paper-heavy and prone to human error. Every technical question I had was met with high-level competence."
— Christoph F., Managing Director, Canova GmbH

Score every candidate process from 1 (low) to 5 (high) on five questions. Work through the highest totals first. Hold back anything that scores 1 on data access or 5 on risk until that is fixed.

Priority = Volume + Rule clarity + Data access + Value + (6 − Risk). The maximum is 25.

How to read it:

1. 20–25: a strong first candidate. Write the spec.
2. 14–19: worth doing once the first system is live, or with a standard tool.
3. Below 14: wait, fix the data first, or keep it manual.

Illustration, not client data: supplier invoice entry often scores high on volume, rule clarity and data access. Quote preparation scores high on value but needs a human approval step. A company-wide chatbot often scores low on rule clarity and value until the underlying knowledge is cleaned up.

AI agents are one possible answer on a roadmap, not the starting point. The roadmap ranks processes first, then picks the simplest thing that works: a rule-based automation, an off-the-shelf tool, or an agent that drafts, routes and files work inside the tools you already use, with a human approving anything that leaves the company. Common starting points Made By Agents ships include quote automation, OCR and document pipelines, knowledge bases, dashboards and multi-agent systems. For ongoing delivery across several initiatives there is the AI Partner retainer, at a fixed monthly rate that you can pause any month.

1. Fixed price and scope in writing before work starts.
2. Named deliverables: a ranked list and a spec, not a slide deck.
3. An honest "don't build" option. The audit should be allowed to say "wait" or "buy a tool".
4. Build-vs-buy neutrality. Ask whether they resell software.
5. Practitioners. The person doing the audit should have shipped systems like the ones they recommend.
6. A credit against implementation, without lock-in. The roadmap should be usable by any competent team.

How much does an AI audit cost?
It depends which audit you mean. An opportunity audit or roadmap costs roughly €1,500–€4,000 as a one-day workshop and €6,000–€15,000 as a multi-week analysis in Germany. ISO/IEC 42001 certification audit fees range from about $5,000 to $65,000 or more.

What is the difference between an opportunity audit and a governance audit?
An opportunity audit asks where AI pays off and what to build first. A governance audit asks whether the AI you already use is managed responsibly, for example against ISO/IEC 42001.

How long does an AI roadmap engagement take?
Published fixed-price offers range from one workshop day to three weeks. The Deep AI Audit starts with a 20-minute call; most audits then take 1–2 weeks, followed by one walkthrough.

What does Made By Agents charge for the Deep AI Audit?
Made By Agents doesn't publish prices. The audit is a fixed fee, quoted after a free Discovery Call once the scope is clear. If implementation with us starts within 90 days, the fee is subtracted from that project.

How much of our team's time does it take?
A 20-minute call, then a few hours with the people and systems that run the process.

Do we have to build with the same firm afterwards?
No. You can build in-house, book sessions, or hire Made By Agents for the first system.

Is the audit remote or on-site?
Remote by default. On-site is possible if you cover the travel costs.

Where do AI agents fit into an AI roadmap?
Agents are one option. The roadmap ranks processes first, and many are better served by simple automation or a standard tool.

Book a free 20-minute Discovery Call. We'll tell you whether a Deep AI Audit is the right next step, or whether something smaller, like a consulting session, is enough.

Last reviewed: 2026-10-31 by Tobias Wupperfeld.