Skip to content
D K
EU AI Act: Not High Risk Q3

Policy Document Agent

One source of truth for every HR policy - always current, always accessible.

Manages HR policies: creation, versioning, approval, publication, and enquiry responses - always the current version available.

Analyse your process
Airbus Volkswagen Shell Renault Evonik Vattenfall Philips KPMG

Versioning rule, content extraction via AI, inquiry routing

The agent manages policies with a high H-share in creation: HR authors determine content and approval, the agent structures versions deterministically, extracts via AI content changes between versions and answers inquiries rule-based with the version valid on the request date.

Outcome: According to industry surveys, around 60 percent of HR policies in mid-market companies are not versioned and not centrally findable; with up to 80 active works agreements, legal uncertainty regularly arises between managers.

57% Rules Engine
29% AI Agent
14% Human

The core is not the document but the question which version applied on day X:

Forty to eighty policies, one rule always violated

A company with 1,500 employees typically maintains between 40 and 80 internal policies: travel expense rules, remote work policy, company car policy, data protection policy, code of conduct, working time models, training policy, anti-discrimination policy - and the list grows with every legal change and every new works council agreement. Every single policy has to be current, findable, understood, and demonstrably acknowledged.

In practice, at least one of these four conditions is almost always broken.

The regulatory operating system - and why it is offline

HR policies are not an administrative sideshow. They are the regulatory operating system of an organisation. They define what is allowed and what is not, what entitlements exist, what obligations apply. Employment tribunals, tax auditors, and data protection authorities measure companies against what is in these documents - and against whether the workforce demonstrably knew about them.

The problem is not that policies are missing. The problem is that the system managing them is a decade behind. A SharePoint folder with 200 files, 40 of them named “final_v3_NEW”. An email from last November with the subject “updated travel policy” - that no one opened. A works agreement on remote work that should have been revised after the last legal change but is stuck in the review loop between legal, the works council, and senior leadership.

The result: HR answers the same questions over and over. Employees act on outdated rules and do not know it. And when a dispute arises, the proof that the current version was communicated is missing.

Three problems that reinforce each other

Version chaos

In most organisations, policies exist in multiple versions simultaneously. The HR department works with the current one, managers have an older one on a local drive, and the intranet shows a third whose validity status is unclear. For a remote work policy that has been amended four times since 2020, this is not a theoretical risk - it is the normal state.

In many smaller organisations, employee handbooks remain unchanged in circulation for years. For mid-sized companies with more complex rule sets, the problem worsens because each policy has dependencies on other policies. A change to the working time rules can affect travel expense policy, overtime rules, and on-call agreements simultaneously.

The acknowledgement gap

A significant share of employees have never read the current policy handbook, particularly in younger cohorts. HR organisations widely struggle just to get employees to take policies into account. A relevant portion does not even know where the documents are stored.

That would be a pure communication problem if acknowledgement did not have legal meaning. But it does. Before employment tribunals and in regulatory audits, what counts is the proof that employees received, read, and understood the policy. Without that proof, the policy is worthless in a dispute - no matter how carefully it was drafted.

Policy written            != Policy communicated
Policy communicated       != Policy read
Policy read               != Acknowledgement documented
Acknowledgement documented =  Policy enforceable

The approval bottleneck

Every policy change in a company with worker representation goes through at least four stations: subject-matter department drafts, legal reviews, works council is involved, leadership approves. Depending on the topic, the data protection officer, safety officer, or external advisors are added.

Works constitution law differentiates by intensity of participation. A remote work policy alone touches rules on workplace order, working time, health protection, and mobile work. Each can trigger its own negotiation process.

Without workflow orchestration, a policy update takes three to six months. During that time, the old version applies - or none at all, if the old one has already been withdrawn. New laws (EU Pay Transparency Directive, NIS-2 transposition, AI Act) keep increasing the update pressure. In Europe, cascading EU directive implementations produce similar patterns to the 15 US states that introduced new employment laws on 1 January 2026 alone.

What the agent actually changes

The Policy Document Agent addresses all three problems simultaneously, because they share the same structural origin: missing versioning, missing process control, missing follow-up.

Versioning with validity periods. Every policy gets a unique version number, an effective date, and an optional expiry. Old versions are archived, not deleted. For every query, the system delivers the version that was valid at the time of the query - not the newest, but the one that applies to the specific case. An employee asking about their February travel expense receives the policy that was valid in February.

Controlled approval processes. A draft for a new or changed policy runs through a configured workflow. Which stations are involved is determined by a rule set based on policy type and participation rights. For topics subject to co-determination, worker representative involvement is scheduled automatically. Deadlines are tracked. Delays become visible before they turn critical.

Verifiable acknowledgement. Once a new policy is approved, it is distributed to the relevant employee groups - not as an email attachment, but with read confirmation and timestamp. Anyone who has not yet confirmed gets a reminder. HR sees the current confirmation status at any time. Before an employment tribunal, a complete record of acknowledgement exists.

When employees ask

The fourth function is the most visible: answering policy questions. An employee asks whether special leave applies for a local move. The agent finds the relevant section in the current version of the agreement, delivers the answer with source reference (section, version, effective date), and marks the question as factual or interpretive. Factual questions are answered directly. Interpretive questions are routed to HR - with the relevant context, so the answer comes faster.

The distinction matters. “How many days of special leave am I entitled to for a move?” is a factual question - the answer is in the works agreement. “Does that also apply if I only move within the same city?” can be an interpretive question that depends on the specific wording. The agent recognises the difference and acts accordingly.

This not only reduces the volume of queries to HR. It changes the quality of the remaining ones. Instead of answering 30 identical questions about remote work quotas, the HR department deals with the three cases that genuinely require judgement.

Infrastructure that reaches beyond policies

Two of the three core components of this agent - versioning engine and approval workflow - are generic infrastructure. Every agent in the Decision Layer that applies rule sets needs versioned documents with validity periods. Every agent that orchestrates multi-level approvals needs a workflow engine with deadline tracking.

The Compliance Monitoring Agent checks operational data against rule sets - these rule sets must be versioned. The Works Council Coordination Agent manages participation processes - these processes follow the same approval logic. The Audit Agent needs proof - the acknowledgement tracking provides it.

Starting with the Policy Document Agent does not just rebuild policy management. It lays the infrastructure for every rule-based decision a later agent in the ecosystem will make. The versioning engine is reused. The approval logic is reused. The audit trail pattern is reused.

When this pays off

The calculation has two sides. The visible one: HR hours no longer spent on standard questions. At 1,500 employees and 20 policy queries per week, with 70 percent answerable automatically, that is 14 fewer queries - at 15 minutes per query, around 180 hours per year.

The less visible but larger one: the risk that does not materialise. An employment tribunal lost because acknowledgement of a policy cannot be proved. A data protection audit that finds the data protection policy has not been updated for two years. A works council challenge because participation in a policy change was not documented.

These risks cannot be put into a spreadsheet. But every Head of HR who has ever lost a proceeding for lack of proof knows the number. (US: similar documentation discipline now applies under state pay-transparency, sick-leave, and AI-hiring laws that multiply the US patchwork of employer obligations.)

Micro-Decision Table

Who decides in this agent?

7 decision steps, split by decider

57%(4/7)
Rules Engine
deterministic
29%(2/7)
AI Agent
model-based with confidence
14%(1/7)
Human
explicitly assigned
Human
Rules Engine
AI Agent
Each row is a decision. Expand to see the decision record and whether it can be challenged.
Receive policy draft or update Classify as new policy, revision, or retirement Rules Engine

Classification based on policy metadata and change type

Decision Record

Rule ID and version number
Input data that triggered the rule
Calculation result and applied formula

Challengeable: Yes - rule application verifiable. Objection possible for incorrect data or wrong rule version.

Route for review Send to required reviewers (legal, HR, works council) Rules Engine

Review routing rules per policy type and change scope

Decision Record

Rule ID and version number
Input data that triggered the rule
Calculation result and applied formula

Challengeable: Yes - rule application verifiable. Objection possible for incorrect data or wrong rule version.

Collect review feedback Aggregate reviewer comments and change requests AI Agent

Automated feedback collection and consolidation

Decision Record

Model version and confidence score
Input data and classification result
Decision rationale (explainability)
Audit trail with full traceability

Challengeable: Yes - fully documented, reviewable by humans, objection via formal process.

Route for approval Send reviewed policy through approval chain Rules Engine

Approval chain per policy type and organisational level

Decision Record

Rule ID and version number
Input data that triggered the rule
Calculation result and applied formula

Challengeable: Yes - rule application verifiable. Objection possible for incorrect data or wrong rule version.

Approve policy Authorise publication of new or updated policy Human

Human approval with legal and governance responsibility

Decision Record

Decider ID and role
Decision rationale
Timestamp and context

Challengeable: Yes - via manager, works council, or formal objection process.

Publish and notify Make policy live and notify affected employees AI Agent

Automated publication with targeted notification

Decision Record

Model version and confidence score
Input data and classification result
Decision rationale (explainability)
Audit trail with full traceability

Challengeable: Yes - fully documented, reviewable by humans, objection via formal process.

Track acknowledgements Monitor employee read-and-acknowledge completion Rules Engine

Completion tracking with deadline escalation

Decision Record

Rule ID and version number
Input data that triggered the rule
Calculation result and applied formula

Challengeable: Yes - rule application verifiable. Objection possible for incorrect data or wrong rule version.

Decision Record and Right to Challenge

Every decision this agent makes or prepares is documented in a complete decision record. Affected employees can review, understand, and challenge every individual decision.

Which rule in which version was applied?
What data was the decision based on?
Who (human, rules engine, or AI) decided - and why?
How can the affected person file an objection?
How the Decision Layer enforces this architecturally →

Does this agent fit your process?

We analyse your specific HR process and show how this agent fits into your system landscape. 30 minutes, no preparation needed.

Analyse your process

Governance Notes

EU AI Act: Not High Risk
Not classified as high-risk under the EU AI Act - the agent manages documents without employment-affecting decisions. Works council co-determination rights apply to many HR policy changes (particularly those affecting working conditions). The agent's approval workflow must include works council consultation where required. GDPR applies to the acknowledgement tracking (recording which employees read which policies).

Assessment

Agent Readiness 68-75%
Governance Complexity 44-51%
Economic Impact 46-53%
Lighthouse Effect 31-38%
Implementation Complexity 34-41%
Transaction Volume Weekly

Prerequisites

  • Policy document management system with version control
  • Policy review and approval workflow infrastructure
  • Publication channels (intranet, employee portal)
  • Acknowledgement tracking capability
  • Works council consultation process for policy changes
  • Legal review process for employment-related policies

Infrastructure Contribution

The Policy Document Agent builds the authoritative policy repository that the Employee Self-Service Agent, Compliance Training Agent, and Compliance Monitoring Agent all reference. Without a single source of truth for policies, every downstream agent that answers policy questions operates on potentially outdated information. Builds Decision Logging and Audit Trail used by the Decision Layer for traceability and challengeability of every decision.

What this assessment contains: 9 slides for your leadership team

Personalised with your numbers. Generated in 2 minutes directly in your browser. No upload, no login.

  1. 1

    Title slide - Process name, decision points, automation potential

  2. 2

    Executive summary - FTE freed, cost per transaction before/after, break-even date, cost of waiting

  3. 3

    Current state - Transaction volume, error costs, growth scenario with FTE comparison

  4. 4

    Solution architecture - Human - rules engine - AI agent with specific decision points

  5. 5

    Governance - EU AI Act, works council, audit trail - with traffic light status

  6. 6

    Risk analysis - 5 risks with likelihood, impact and mitigation

  7. 7

    Roadmap - 3-phase plan with concrete calendar dates and Go/No-Go

  8. 8

    Business case - 3-scenario comparison (do nothing/hire/automate) plus 3×3 sensitivity matrix

  9. 9

    Discussion proposal - Concrete next steps with timeline and responsibilities

Includes: 3-scenario comparison

Do nothing vs. new hire vs. automation - with your salary level, your error rate and your growth plan. The one slide your CFO wants to see first.

Show calculation methodology

Hourly rate: Annual salary (your input) × 1.3 employer burden ÷ 1,720 annual work hours

Savings: Transactions × 12 × automation rate × minutes/transaction × hourly rate × economic factor

Quality ROI: Error reduction × transactions × 12 × EUR 260/error (APQC Open Standards Benchmarking)

FTE: Saved hours ÷ 1,720 annual work hours

Break-Even: Benchmark investment ÷ monthly combined savings (efficiency + quality)

New hire: Annual salary × 1.3 + EUR 12,000 recruiting per FTE

All data stays in your browser. Nothing is transmitted to any server.

Policy Document Agent

Initial assessment for your leadership team

A thorough initial assessment in 2 minutes - with your numbers, your risk profile and industry benchmarks. No vendor logo, no sales pitch.

30K120K
1%15%

All data stays in your browser. Nothing is transmitted.

Frequently Asked Questions

How does the agent handle policies that require works council agreement?

The agent's approval workflow includes works council consultation as a required step for policy types that fall under co-determination rights. The policy cannot be published until works council agreement is documented.

What happens when an employee accesses an outdated policy version?

The agent maintains version control with clear effective dates. Employees always see the current version by default. Historical versions are retained for audit purposes but clearly marked as superseded.

What Happens Next?

1

30 minutes

Initial call

We analyse your process and identify the optimal starting point.

2

1 week

Discover

Mapping your decision logic. Rule sets documented, Decision Layer designed.

3

3-4 weeks

Build

Production agent in your infrastructure. Governance, audit trail, cert-ready from day 1.

4

12-18 months

Self-sufficient

Full access to source code, prompts and rule versions. No vendor lock-in.

Implement This Agent?

We assess your process landscape and show how this agent fits into your infrastructure.