Your browser does not support JavaScript! Please enable the settings.
↑

EU AI Act Dates After the Digital Simplification Report: What UK Product Teams Should Plan Around

The EU AI Act's high-risk obligations are still shown for 2 December 2027, but the EU's simplification agenda has put dates under discussion. This guide separates what is confirmed from what is proposed and sets out a plan that works either way.
November 20, 2026
Innovify Editorial
published on
November 20, 2026
EU AI Act Dates After the Digital Simplification Report: What UK Product Teams Should Plan Around

EU AI Act Dates After the Digital Simplification Report: What UK Product Teams Should Plan Around

UK product teams selling into the EU have been told two things about the EU AI Act this autumn. First, that the high-risk obligations, the ones that matter most for credit scoring and similar financial use cases, apply from 2 December 2027. Second, that the EU is simplifying its digital rulebook and some of that may change. Both are true, and together they create a planning problem: budget for a deadline that might move, or wait and risk being late. On 4 November 2026 the European Commission published its digital simplification report, yet its AI Act page still shows 2 December 2027 for high-risk obligations. The detailed omnibus text needs a careful read before anyone changes advice. This guide separates what is confirmed from what is proposed, and sets out a plan that works under either outcome.

This is a technology partner's reading of public material, not legal advice. Check the Commission's own pages and take advice on your own position.

Why UK teams cannot ignore the EU AI Act

The UK has not adopted the EU AI Act. Its approach relies on existing regulators, such as the FCA and the ICO, applying their current rules to AI. That does not take UK firms out of scope. The Act can apply to organisations outside the EU when they place AI systems on the EU market or when the output of their systems is used in the EU.

For a UK fintech, lender, insurer or platform with EU customers, the question is usually not whether the Act exists but which of its obligations apply to which of its systems, and when. That is why the dates matter. A credit-scoring model used for EU applicants, for example, falls into the category the Act treats as high-risk, and that category carries the heaviest obligations.

Interest is rising. UK monthly searches for "eu ai act" reached about 8,100 in September 2026, with a high of about 9,900 in August, according to Ubersuggest estimates used in our market intelligence. Governance terms also carry high paid competition, which suggests advertisers see commercial value in the topic.

What is confirmed today

The safest planning basis is what the European Commission states on its own pages.

The Act is law. It entered into force in 2024 and applies in stages, not all at once.

High-risk obligations are shown for 2 December 2027. As of the Commission's AI regulatory framework page, read in November 2026, that is the date given for high-risk obligations. Our market intelligence found no change to it on the Commission page after the simplification report.

Harmonised standards are arriving. Technical standards that help firms show compliance are being developed and published. They matter because following a harmonised standard can support a presumption of conformity for the requirements it covers.

What "high-risk" means for financial services

The Act sorts AI systems by risk. High-risk systems carry obligations covering risk management, data quality and governance, technical documentation, record-keeping, transparency to users, human oversight, and accuracy, robustness and security. Systems used to evaluate creditworthiness or establish credit scores of individuals are among the use cases the Act lists as high-risk. Some insurance pricing and underwriting uses are also covered.

For many UK fintechs, that list is the main reason the Act reaches them.

What is proposed, and not yet settled

The EU's digital simplification agenda aims to reduce compliance burden across several digital laws. Proposals of this kind can adjust scope, timing or detail. The point for product teams is narrow: until the final text is read and adopted, a proposal is not a change.

That leaves three realistic scenarios.

  1. Dates hold. High-risk obligations apply from 2 December 2027 as currently stated.
  2. Dates move later. Some obligations are deferred, for example to allow standards to mature.
  3. Scope or detail changes. Dates hold or move, but what is required is simplified or clarified for some categories.

No one can tell you today which will happen. What you can do is plan so that your work is useful in all three.

A plan that works under every scenario

The common mistake is to treat EU AI Act readiness as a single project with a single deadline. A better approach splits the work into what is valuable regardless of the final dates, and what is worth timing to the confirmed deadline.

No-regret work: start now

Some work is valuable even if the Act's dates move, because UK regulators, bank partners, auditors and enterprise customers already expect it.

  • An AI system inventory. List every model and AI-driven feature, what it does, what data it uses and who it affects.
  • A risk classification per system. For each, decide whether it could fall into a high-risk category if used for EU customers.
  • Documentation by design. Record design choices, training data sources, testing and known limitations as you build, rather than reconstructing them later.
  • Human oversight points. Decide where a person reviews, approves or can override an AI-driven outcome.
  • Logging. Keep records of inputs, outputs and decisions in a form you can retrieve and explain.

None of these depends on the final EU text. The FCA's October 2026 document on AI good practice and the ICO's call for evidence on agentic AI point in the same direction: be able to show what your AI does, who owns it and how it is controlled.

Timed work: plan to the confirmed date, review quarterly

Other work is specific to the Act and worth sequencing against the current date, with a scheduled review.

  • Conformity assessment preparation for systems that are high-risk in scope.
  • Quality management system elements specific to the Act's requirements.
  • Alignment with harmonised standards as they are published.
  • Contract changes with EU customers and suppliers.

Set the plan against 2 December 2027, and review it each quarter against the Commission's pages. If the date moves, you release budget. If it holds, you are on track.

What this means for different UK teams

Lenders and credit platforms with EU applicants are the most directly affected, because creditworthiness assessment is a listed high-risk use case. Start with your scoring and decisioning models.

Insurers and insurtechs should check whether their pricing or underwriting models fall into covered categories for EU customers.

Embedded finance and BaaS platforms should map where AI decisions sit in the chain: with the platform, the regulated partner or the brand. Responsibility follows the role, so the contracts matter.

AI-native product teams should check whether they are providers or deployers of the systems they use, as obligations differ. A team building on a third-party model still has its own duties as a deployer.

A quarterly review routine

Because the dates may move, readiness needs a light, regular review rather than a one-off project plan. A quarterly routine works well for most product teams.

Check the source. Read the European Commission's AI Act pages and any adopted amendments. Note whether the high-risk date or scope has changed. Do not rely on secondary summaries alone.

Refresh the inventory. New models and AI features appear quickly. Add them, retire what is no longer used and re-check the risk classification of anything whose purpose has changed.

Check the evidence. Pick two or three systems at random and test whether the documentation, oversight records and logs would stand up to a request from a customer, partner or regulator. Gaps found in a quarterly check are cheap to fix. Gaps found during an audit are not.

Re-plan timed work. If the confirmed date has moved, adjust the sequence and budget for conformity assessment and standards alignment. If it has not, confirm you are on track.

Report briefly. Give leadership a one-page update: what changed, what is on track and what decision, if any, they need to make.

Questions your board will ask

Boards and risk committees tend to ask the same questions. It helps to have answers ready.

  • Which of our AI systems could be high-risk if used for EU customers?
  • What is our confirmed deadline, and what would change if it moved?
  • Who owns EU AI Act readiness, and how is progress reported?
  • What are we doing now that also satisfies UK regulators and partners?
  • What would it cost to be late, and what would it cost to be early?

Who should own readiness

EU AI Act readiness touches product, engineering, data, risk, compliance and legal. Give one person end-to-end ownership, with a small group from each function meeting each quarter. Without a single owner, inventories go stale and timed work slips.

Common planning mistakes

  • Waiting for certainty. The no-regret work is useful under every scenario and is easier the earlier it starts.
  • Treating the UK as exempt. The Act can apply to UK firms serving EU users.
  • Confusing a proposal with a change. Plan to adopted text, not headlines.
  • Running EU, FCA and ICO work separately. Much of the evidence overlaps. One inventory and one documentation habit can serve all three.

Where Innovify fits

Innovify is a London-based technology partner with deep expertise in banking, financial services, fintech and insurance, and we design compliance in from day one. Through our AI Consulting work we help teams build an AI system inventory, classify risk and separate no-regret work from date-dependent work. Our AI / ML Development teams then build documentation, oversight and logging into the systems themselves, so readiness is part of the product rather than a separate compliance exercise.

FAQ

When do the EU AI Act high-risk rules apply?

As of November 2026, the European Commission's AI regulatory framework page shows 2 December 2027 for high-risk obligations. The EU's digital simplification agenda could affect timing or detail, so firms should check the Commission's own pages and plan against the confirmed date with a regular review.

Does the EU AI Act apply to UK companies?

It can. The Act applies to organisations that place AI systems on the EU market or whose system outputs are used in the EU, even if they are based outside it. A UK lender scoring EU applicants, for example, may be in scope. The UK itself has not adopted the Act.

Is credit scoring high-risk under the EU AI Act?

Yes. AI systems used to evaluate the creditworthiness of individuals or establish their credit scores are listed as high-risk use cases. They carry the Act's heaviest obligations, including risk management, documentation, logging, human oversight and accuracy requirements.

Should we wait for the EU's simplification proposals before preparing?

No. Much of the preparation, such as an AI inventory, risk classification, documentation, oversight and logging, is valuable regardless of final dates and is already expected by UK regulators and partners. Time only the Act-specific work, such as conformity assessment, to the confirmed date.

How does the EU AI Act relate to UK regulation?

The UK relies on existing regulators applying their rules to AI rather than a single AI law. The FCA has published AI good-practice material and the ICO has consulted on agentic AI. Much of the evidence they expect overlaps with EU AI Act requirements, so one programme can serve both.

Conclusion

The EU AI Act's high-risk date is still shown as 2 December 2027, and the simplification agenda has made people unsure whether it will hold. The answer is not to wait. Do the work that is valuable under every scenario now: inventory, classification, documentation, oversight and logging. Time the Act-specific work to the confirmed date, and review it every quarter. That way a moving date becomes a budgeting question, not a compliance crisis.

Related reading

Connect With Us

Let’s Build Something Exceptional

Innovify transforms ideas into scalable, impactful products by integrating user centered design, AI-native engineering, data-back growth and lean agile methodologies. Our close client collaboration complemented with on-shore and nearshore delivery creates speed, clarity, and success at scale.
Discuss your project