30 October 2026 is ERA sexual-harassment reform. FluencyRank now ships era_harassment_ai practice alongside named_policy_ai Duty drills. Practice evidence for teams — not “satisfies all reasonable steps.”
Rafał RadziszewskiFounder, FluencyRank
UK employers face two different pressures that sales teams often blur. Consumer Duty embedding and Conduct-style judgment when AI drafts reach customers. And — separately — Employment Rights Act reform from 30 October 2026 raising the sexual-harassment prevention duty to “all reasonable steps,” including third-party harassment.
FluencyRank ships a dual UK stack on the same engine: era_harassment_ai for escalate / override / third-party judgment with AI still on the card, and named_policy_ai for Consumer Duty / Conduct-style named-policy drills. Champions get practice evidence in the Evidence Pack — an engagement signal for leaders, not a compliance certificate and not legal advice.
Allowed claim: workplace practice ready before 30 October 2026. Forbidden: satisfies ERA / “all reasonable steps,” FCA, Consumer Duty, or Art. 4 — and not a Skillcast replacement.
Calendar — three clocks, one dual stack
- Consumer Duty: in force since 2023. 2025–26 is embedding and scrutiny — covered in practice by named_policy_ai, not by the ERA date.
- EU AI Act Article 4 literacy: proportionate measures; often-cited enforcement windows around early August 2026 for EU-facing firms — separate from UK employment law.
- 30 October 2026: ERA sexual-harassment reform (“all reasonable steps,” including third-party). Covered in practice by era_harassment_ai — practice readiness for teams, not legal satisfaction.
Do not say “FCA-ready by 30 Oct.” Do say “practice ready before 30 Oct for ERA-related workplace judgment with AI on the card,” plus the non-claim line above. Confirm any regulatory file with counsel.
What continuous practice looks like
- ERA pack: AI drafts that demean, soft-pedal reports, stereotype, or laugh off third-party abuse — escalate and override
- Named-policy pack: Consumer Duty / vulnerable customer / override-before-send when AI drafts go to customers
- First answer scored Correct / Not quite; iteration when judgment matters
- Evidence Pack records who practiced which themes — decision support, not certification
What to avoid
- “Satisfies all reasonable steps” or ERA / FCA certification language
- One-off LMS completion sold as Duty or ERA satisfaction
- Replacing Skillcast / conduct LMS modules — different category
- Dropping AI from the scenario and still calling it FluencyRank
Try the ERA harassment practice demo (HR) or the named-policy Duty demo (finance), or request a B2B evaluation. Organizations remain responsible for their AI and employment policies.
Sources
Related insights
- Article 4 wants documented effort by role — not theater certificates.
Commission guidance frames AI literacy as supporting development with measures by role and context — not one “sufficient” score for everyone. Theater certificates still miss the point.
- Continuous AI readiness is a loop, not a workshop.
Seats, Friday workshops, and a once-a-year quiz decay. Readiness holds when teams baseline, practice in real work, keep evidence, and improve on a cadence.
About the author

Rafał Radziszewski · Founder, FluencyRank
Rafał Radziszewski is the founder of FluencyRank. He is a Senior Director of Engineering and Poland Site Lead with 17+ years in commercial software and nearly a decade leading engineering teams — including FinTech, pharma, industrial enterprise, consulting/outsourcing delivery, and retail & corporate banking at Bank Millennium. He focuses on practical AI enablement for workplace teams — building fluency through daily practice, not tool rollouts alone.