Transparency marking deadline: 121 days · 2 December 2026 Diagnostic delivered in 1 week, full conformity path in 6 weeks

Get your AI compliant with EU law in 6 weeks.

For European companies running AI in credit, hiring, healthcare, identity or critical infrastructure: the 2 August 2026 duties are already in force and Annex III conformity binds on 2 December 2027. We deliver the full Conformity Assessment package, every document the regulator demands, signed and ready to defend, without slowing your roadmap.

Book a Readiness Diagnostic · €1,000 Schedule a free 30-minute call Check your EU AI Act risk free, 90 seconds

How the Diagnostic works. Automated. You answer a short intake about your AI systems and we produce a full readiness report: every system classified by risk, every EU AI Act rule that applies listed with its deadline, every gap counted, and the top 3-5 actions ranked by priority. If you continue with us, the Diagnostic price is credited in full toward any next step. If you don't, you keep the report.

Trusted by leaders. Backed by AI safety experts from
What this law is

The EU AI Act, explained in 60 seconds.

The European Union passed a law called the EU AI Act. Since 2 August 2026 the prohibitions, the Article 50 transparency duties, the general-purpose AI model obligations and the regulators' enforcement powers have all been binding, and the Commission's own AI Act service desk confirms those dates. If your company uses AI to make decisions about credit, hiring, healthcare, identity, education or critical infrastructure, your systems are Annex III high-risk and their conformity obligations bind on 2 December 2027. That is 16 months to produce evidence that takes most teams the better part of a year, on systems already in production today. The penalty for failing high-risk obligations reaches €15 million or 3% of global annual revenue, whichever is higher, and for SMEs and start-ups whichever is lower (Article 99).

Note: obligations for Annex III high-risk AI systems were extended to 2 December 2027 by the EU Digital Omnibus Act. The 2 August 2026 date governs prohibitions, transparency (Art. 50), GPAI rules, enforcement powers, and the Art. 49 registration database, all of which apply from that date.

Proving it means producing a specific set of documents the European Commission has named: a risk management procedure, a data governance file, a human oversight design, a technical specification, and a quality management plan. Then signing a declaration that the documents are accurate. Most companies in scope have produced none of them.

The law applies to your company even if it is not based in Europe. If your customers are in the EU, or your AI affects EU users, or your output is used in the EU, you are in scope. The law spells this out in Article 2.

48.6% have not started.

Deloitte Legal surveyed 500 people responsible for AI in their companies in September 2024. 48.6% had not engaged seriously with implementing the EU AI Act, and only 26.2% had actively begun. Source: Deloitte Legal, AI Act Survey 2024.

The part that is not a future deadline

One duty on this list is already late.

Every date above is a date to prepare for. Article 4 is not. The AI literacy duty has bound both providers and deployers since 2 February 2025, and on 24 July 2026 Regulation (EU) 2026/1744, the Digital Omnibus, replaced its wording with a version that applies from 27 July 2026. The Omnibus moved the Annex III high-risk conformity deadline to December 2027. It did not move this one.

Here is the whole duty, in the words the law now uses. Providers and deployers "shall take measures to support the development of AI literacy of their staff and other persons dealing with the operation and use of AI systems on their behalf", and the same article adds that this "does not require providers or deployers to guarantee any specific level of AI literacy of any individual".

Two things follow, and both matter more than they look. First, it lands on deployers, which the Act defines as anyone using an AI system under their own authority for professional purposes. That is not the narrow set of companies building high-risk systems. If your team has been using AI assistants for the past 18 months, that is you. Second, it is a duty to take measures, not to certify anybody, so it is genuinely finishable. There is no prescribed course and no certificate.

We will be straight with you about the stick, because you will hear otherwise: Article 4 is not named in Article 99, so no administrative fine attaches to it on its own. The reason to close it is not a fine. It is that a regulator asking about anything else on this page will ask what measures you took, and "none, we were waiting for December 2027" is an answer given by a company that was told and did nothing.

AI Literacy Baseline, €500, delivered in 5 working days.

A documented set of measures, an inventory of the AI systems your staff already use, and a training plan. Credited in full against the €1,000 Diagnostic, so it costs nothing to start here and continue.

Start the AI Literacy Baseline · €500

Cost of inaction vs cost of compliance

See your numbers side by side.

Slide your annual revenue. Compare what non-compliance could cost you against what 6 weeks with us costs.

Annual revenue €25,000,000
€0€100M€500M€1B
Cost of inaction €15,000,000 Maximum Article 99 fine for failing high-risk obligations
60%of your annual revenue
7.2 monthsof revenue, gone
Cost with Better Societies €25,000 Done-For-You · replaces a €312,000+ Big 4 quote
0.07%of your annual revenue
0.3 daysof revenue
Net protection vs cost of inaction €14,982,000 832× return on every euro invested with us

If those numbers make sense, find your exact exposure in one week with the Diagnostic →

Where we fit

Three ways companies handle this. Only one is right for most.

Compliance software gives you a dashboard. Big 4 Consulting gives you an associate-led retainer. We sit deliberately in the gap.

Compliance software

Vanta · Drata · Holistic AI · Credo
€25,000per year, recurring

A dashboard that tracks which controls your team has and has not built. Useful for ongoing monitoring. But it does not write your technical file, your risk assessment, or your oversight design. You still have to find a human to do that work.

Better Societies

Done-for-you, fixed fee
€25,000Done-For-You · fixed fee · not recurring

We produce every document the law requires, plus a full AI governance framework, vendor questionnaire, registration support, a 12-month roadmap, one 90-minute call with Adrien, and 30 days of email Q&A. Prior diagnostic fee credited in full.

Big 4 advisory

PwC · Deloitte · EY · KPMG
€300,000+3 to 6 months

A team of associates billed by the hour, scoped twice before you see an invoice, three to six months to complete. Built for Fortune 500 budgets. Overkill for most companies and far slower than the deadline allows.

Pricing

Pick the engagement that matches where you are.

01 · Diagnostic

Readiness Diagnostic

Start here if you are not sure whether the law applies to you, or which of your AI systems it applies to.

€1,000
Fixed · 1 week

Fee credited in full against every later step

  • An inventory of every AI system in your company, classified against the law
  • A clear answer on which of your systems are high-risk and which are not
  • A gap analysis of what compliance work each high-risk system needs
  • A prioritised roadmap with cost and time estimates for each gap
  • A 30-minute walkthrough call with Adrien Harrison to review the report

If you continue to any next step, the €1,000 Diagnostic fee is credited in full. At the Done-For-You tier, it's effectively free.

Book the Diagnostic

Not ready to buy? Map every obligation free in 10 minutes →

02 · Done-For-You

Done-For-You Engagement

If you know the law applies and you want everything handled, documentation, governance, registration, roadmap, and direct access to Adrien.

€312,000€25,000
Fixed · end-to-end · one price for the whole engagement
The struck-through figure is our own worked estimate of the same scope bought line by line, shown in full further down this page. It is an illustration we built, not a quote anyone has shown us.
  • Full evidence file a regulator or enterprise buyer would ask for
  • Full AI governance framework and vendor/supplier AI questionnaire
  • EU AI Act database registration support (where required)
  • Signed Declaration of Conformity + one-page plain-English conformity summary
  • 12-month compliance roadmap and 6-month reminder calendar
  • One 90-minute call with Adrien Harrison + 30 days of email Q&A

50% on signature · 50% at the week-three milestone, after you've reviewed the inventory, classification, and first documentation drafts. Full refund if scoping reveals the work is out of scope.

Start Done-For-You · €25,000
03 · Enterprise

Enterprise Custom

Multiple AI systems, multiple jurisdictions, or board-level briefings required.

From €30,000
Quoted · 8 to 12 weeks
  • Coordination across four or more AI systems in a single engagement
  • Multi-jurisdiction mapping (EU, UK, US states, APAC where relevant)
  • Executive briefings for C-suite and board, prepared and delivered
  • Direct liaison with your in-house counsel or external law firm
  • Eligible for the Conformity Watch retainer for ongoing monitoring (priced separately)

Quote returned within 2 business days of the scoping call.

Request a scoping call
Inside the €25,000 Done-For-You engagement

What you receive, and what each piece would cost from a Big 4 firm.

The Done-For-You engagement bundles six core deliverables and three included extras, plus a full AI governance framework, vendor questionnaire, registration support, a 12-month roadmap, one 90-minute call with Adrien, and 30 days of email Q&A. Big 4 firms charge for each one separately, on hourly retainers, with associate teams that re-scope twice before you see a final invoice.

Risk Management System (Article 9)Risk register, mitigation measures, residual risk scoring per high-risk system
€60,000+at Big 4
Data Governance File (Article 10)Training data documentation, bias monitoring, provenance, retention policy
€45,000+at Big 4
Human Oversight Design (Article 14)Where humans intervene, override authority, escalation paths, logging
€35,000+at Big 4
Technical Documentation (Annex IV)Architecture, model cards, evaluation, deployment, 80+ page submission file
€90,000+at Big 4
Quality Management System (Article 17)Operations, testing, deployment, post-market monitoring, incident response
€40,000+at Big 4
Declaration of Conformity (Article 47)Final signed declaration, CE marking instructions, EU database registration guide
€20,000+at Big 4
Included extra · 12-month auditor-acceptance guaranteeIf your auditor rejects our docs in the first 12 months for AI Act reasons, we revise at no extra cost
€8,000typical
Included extra · Customer vendor questionnaire pre-fillsTwo enterprise procurement questionnaires answered with citations
€4,000typical
Included extra · Direct access to Adrien HarrisonFounder-level email and Slack throughout the six weeks, no associate gating
€10,000typical
Our estimate of the same scope, bought line by line€312,000+
€25,000Your investment

How we built that total, so you can check it or disagree with it. Every line above carries a price we set ourselves, and the total is their sum. Big 4 firms do not publish a rate for EU AI Act conformity work, so there is no public list price to quote against and we do not pretend to have seen one. They do publish day rates for consultancy on public procurement frameworks, for example the Deloitte LLP G-Cloud 14 rate card published by the UK government, which is where anyone can see the shape of that market for themselves. Treat our total as a stated assumption, not a finding.

€287,000
The gap between our fixed fee and the line-by-line estimate above, which is €312,000 minus €25,000
6 weeks
Fixed, not an open-ended associate retainer
One fee
No re-scoping cycles and no change orders

Conformity Watch, optional ongoing retainer. The AI Act keeps moving. Harmonised standards finalise through 2026 and 2027. New AI Office guidance lands every quarter. Member states pass implementing acts. Conformity Watch keeps your documentation in lockstep: monthly regulatory-update digests, quarterly documentation reviews and updates, customer and auditor questionnaire-response support, and priority founder access. Pricing scales with your number of high-risk systems. We discuss it on the discovery call, no pressure to take it, no penalty if you don't.

How it works

6 weeks. 6 deliverables. Every week ends with a signed document.

Week 1

Inventory & risk class

We map every AI system in your company against the law's risk categories.

Week 2

Risk management

We document the risks each high-risk system creates and the controls that mitigate them.

Week 3

Oversight & data

We design where humans intervene in your AI, and audit the data you trained it on.

Week 4

Technical file

We produce the multi-section technical specification your auditor or customer reads.

Week 5

Quality management

We document how you build, test, deploy, and update AI systems on an ongoing basis.

Week 6

Conformity & handover

Signed declaration, submission package, sign-off meeting with your team.

Each week ends with a 30-minute review call and a signed deliverable. Adrien Harrison, our founder and CEO, personally reviews and approves every document before it leaves our hands. No associates. No handoffs.

Conformity guarantee

If we don't deliver, you don't pay.

We offer the strongest guarantee on the EU AI Act compliance market. Three commitments, in writing, backed by a money-back clause in the engagement contract.

1 · Delivery on schedule

If we don't deliver every document on the engagement spec sheet by week six, you get a full refund of the second 50% payment. No arguments.

2 · Acceptance by your auditor

If your auditor, notified body, or enterprise customer rejects our documentation in the first twelve months for AI Act compliance reasons, we revise it at no extra cost until accepted.

3 · Out-of-scope refund

If our discovery call reveals the engagement is out of scope for what we deliver, for example, you're a deployer not a provider, we refund your deposit, no questions asked.

Who's signing your documents
Adrien Harrison

Adrien Harrison

Founder & CEO · Better Societies

Forbes-featured serial entrepreneur working in AI safety strategy, based in New York City.

Specialised in AI safety, deep-tech regulation, and responsible-innovation governance. Has worked with founders, investors, researchers, and policymakers from 50+ countries since 2015.

Every Better Societies engagement is led by Adrien personally. Documentation is reviewed and signed by him before delivery. There are no associates, no handoffs, no name partners who appear only in the contract.

Adrien on LinkedIn →
What this looks like

Three demonstrative engagements. Real mechanics, example companies.

Real client engagements are confidential under NDA. The three case studies below walk through what happens when companies matching three common profiles engage Better Societies, with the actual savings calculated against named alternatives.

Fintech · Series B

How a 180-person Amsterdam fintech with three high-risk credit systems would save €280,000 and four months.

Their alternatives: a €380,000 PwC quote over four months, or a Vanta dashboard that doesn't write the technical file. Better Societies delivers the same work for €25,000 Done-For-You, including the governance framework, vendor questionnaire, and a 12-month roadmap, leaving engineering free to keep shipping product.

Read the full case study → HR-tech · Pre-Series A

How a 22-person Berlin candidate-screening startup would unblock a €500,000 enterprise contract in six weeks.

A DAX-listed customer requires AI Act documentation in their procurement form. Series A diligence is asking the same questions. Six weeks of Better Societies work delivers the technical file plus an investor-ready compliance summary, closing both at once.

Read the full case study → Partnership · Law firm

How an 8-attorney Dublin privacy law firm would add €40,000 in recurring revenue without hiring.

Their clients keep asking about the AI Act. Hiring a compliance engineer costs €130,000 per year. Referring to Better Societies pays €5,100 per client across the engagement and the retainer, with twelve hours of partner time per referral.

Read the full case study →
Questions

Five questions we hear most often.

Does this law apply to my company if we are not based in Europe?
Yes, if any of these is true: your AI system is sold or used in the European Union, your customers or their users are in the EU, or the output of your AI is used in the EU. A US fintech serving European borrowers, a UK HR platform with German customers, an Asian healthcare-AI firm with EU clinics, all are in scope. The law specifies this in Article 2.
How do I know if my AI is high-risk?
Two ways. First, if your AI is used as a safety component of a regulated product, medical devices, machinery, vehicles, it is high-risk by default. Second, if your AI makes decisions about credit, hiring, healthcare, identity, education, critical infrastructure, migration, or justice, it is high-risk by category. The Readiness Diagnostic gives you a definitive classification across every AI system in your company in one week.
How are the fines actually calculated?
Three tiers. €35 million or 7% of global revenue for the most serious violations, using AI in prohibited ways. €15 million or 3% of global revenue for failing to meet high-risk obligations, which is what most companies face. €7.5 million or 1% of global revenue for incorrect information to authorities. The fine is the higher of the fixed cap or the percentage, except for SMEs and start-ups, where Article 99(6) reverses it and the lower of the two applies. For a company over €500M revenue, the percentage is what binds.
Why is your fee so much lower than a Big 4 engagement?
Two reasons. First, specialisation: we work on the EU AI Act and nothing else, so the technical documentation, the risk management file, the data governance file and the declaration of conformity all start from templates we already hold rather than being drafted from scratch. Second, methodology: a Big 4 engagement staffs a generalist team billing by the hour, and those firms publish day rates on public procurement frameworks such as the UK government G-Cloud rate cards if you want to see the shape of that market for yourself. Ours is a single fixed fee of €25,000 for a six-week engagement, whatever the hours turn out to be, because Adrien Harrison does the high-judgment work directly. We do one thing. Big 4 does hundreds of things at scale, and you pay for that scale.
What if my AI is not actually high-risk?
Then you have far fewer obligations and the Readiness Diagnostic will tell you that in week one. Most companies have a mix: some systems are high-risk, others are limited-risk (transparency obligations only), others are minimal-risk (basically no obligations). The Diagnostic separates them so you do not over-spend on systems that do not need full conformity work.
Before you commit

Honest disqualification. Better-fit clients save us both time.

This is for you if

  • You build or sell AI that makes decisions about credit, hiring, healthcare, identity, education, infrastructure, migration, or justice
  • You serve EU customers, or have EU users, or your AI output is used in the EU
  • You need the documentation signed and ready for the duties already in force, before your next enterprise deal closes, or before your next funding round
  • You want a fixed price and a fixed timeline, not a Big 4 retainer that re-scopes twice before you see an invoice
  • You have a single decision-maker in the company who can approve scope and sign documents

This is NOT for you if

  • You have not yet built any AI features and want help designing them, that is product work, we don't do it
  • Your AI is purely internal (e.g., grammar checker, code completion) with no EU users or customers, you may not be in scope at all, and the Diagnostic will tell you that
  • You want generic GDPR or SOC 2 dashboard work, see Vanta or Drata for that
  • You want fractional compliance staffing or an ongoing retainer, we do fixed-scope projects
  • You expect to negotiate the deliverable list mid-engagement, our scope is fixed by the spec sheet from day one
Find out where you stand

Two minutes. Your starting step appears immediately.

Tell us what you have so we can tell you what you need. We follow up personally within one business day.

Your details are not shared. Your starting step appears immediately, on this page. We follow up personally within one business day.

Strong fit. Here is how to start.

Based on your answers, the Readiness Diagnostic is built for you. Secure your slot now, or book a free call first. Either way we follow up personally within one business day.

Your compliance path

Step 1 · €1,000 Readiness Diagnostic (1 week): AI system inventory, risk classification, gap analysis, prioritised roadmap. Book the Readiness Diagnostic · €1,000
Step 2 · €2,500 Implementation Accelerator (2 weeks): top gaps closed, technical documentation started. Step 1 fee credited. Start the Implementation Accelerator · €2,500
Step 3 · €5,000 Conformity Sprint (2 weeks): all gaps remediated, full documentation set. Prior fees credited. Start the Conformity Sprint · €5,000
Step 4 · €25,000 Done-For-You (full engagement): governance framework, vendor questionnaire, registration support, 12-month roadmap, one 90-min call with Adrien, 30 days email Q&A. All prior fees credited. Start Done-For-You · €25,000

Each step credits toward the next. The Diagnostic is the intended starting point, and it is the step we scope everything else from.

Book and pay the Diagnostic, €1,000Book a free 30-minute call
Ready?

Two ways to start.

Diagnostic if you know you need the work. Discovery call if you want to talk it through first.

Book the Diagnostic · €1,000 Schedule a free Discovery Call
Diagnostic · €1,000 Free call