Companies that deploy AI to substitute for human sales representatives now operate inside the EU AI Act's high-risk perimeter. Non-compliance can cost up to €35 million or 7% of global annual turnover, whichever is higher. The regulation took effect August 1, 2024, and treats employment-related AI as a category requiring conformity assessments, technical documentation, and human oversight.

What did the EU actually classify as high-risk?

Annex III of the Act lists eight high-risk use cases. Section 4 covers "Employment, workers management and access to self-employment." Within scope: AI used to recruit or select applicants, decide on terms of work, promotion or termination, allocate tasks, or monitor performance.

Sales-replacement deployments — where a buyer meets only an AI agent before contract — fall closer to the "decisions on terms" lane than most vendors acknowledge. Classification turns on use, not on the AI tool's branding.

What does compliance actually require?

A high-risk system must:

• Carry a CE marking and an EU declaration of conformity
• Maintain a risk-management system across its lifecycle
• Log automatically generated events for traceability
• Ship with clear instructions for deployers
• Register in the EU database before market placement

Most obligations for systems already on the market applied from August 2026. New high-risk systems launched after the August 2024 effective date had to comply within 24 months. Companies that built sales-replacement AI in 2023 and shipped it without those controls now face enforcement exposure, not future planning.

Why does the unite.ai argument land?

Titled "Replacing Your Sales Reps with AI Was Always a Risk. The EU Just Proved Why," the piece frames the regulatory shift as confirmation of a pre-existing operational hazard. The argument: sellers measured the cost of headcount, not substitution. The Act's penalty structure — €35M or 7% of turnover, topping the GDPR's 4% cap — recalibrates that math.

A skeptic reads the same text and asks what the Act does not cover. It does not require firms to keep human sales staff. It also does not ban AI in sales. What it does is shift the burden: organizations that replace reps with AI inherit documentation, monitoring, and audit obligations that human-led workflows never carried.

What changes for sales leaders, by team size?

• Under 50 reps: A bot handling first-touch qualification now triggers high-risk duties. The penalty ceiling is the same as for a 5,000-person enterprise, though actual fines typically scale with company size and severity.
• 50–500 reps: Hybrid models — AI for lead scoring, humans for closing — sit in a lower-risk zone, but only if the AI does not autonomously reject candidates or block human follow-up.
• 500+ reps: Enterprise teams absorb compliance workload through existing legal staff. Smaller competitors that automated first and built compliance later carry the higher relative cost.

What is still unsettled

The Act's enforcement record will determine whether the €35M figure is a real ceiling or a paper one. Vendors that sold sales-replacement AI without flagging the high-risk classification may also face downstream liability. The next 18 months of EU member-state penalties will show whether the math changed for sellers — or for the firms that sold them the tools.