Generative AI is making it easier for HR teams to produce policies in minutes. But a new review from BrightHR suggests that speed can come at a cost: an AI-generated absence policy may look comprehensive while leaving out the practical details employers need to apply rules consistently and fairly.
The appeal is obvious. An HR manager can enter a short prompt into an AI assistant and receive a polished absence policy, complete with headings covering reporting requirements, absence monitoring and disciplinary procedures. The problem is that a document that reads professionally is not necessarily a policy that works safely in the workplace.
That distinction is at the centre of BrightHR’s latest examination of AI-generated HR policies. The company’s Head of HR Advisory and Technical Services, Gemma O’Connor, reviewed an AI-written absence management policy created from a simple prompt that instructed the system to state that employees would receive a formal warning after a third absence.
Her conclusion: the AI-generated document covered many of the expected topics at a surface level but lacked operational detail that can become important when managers have to apply the policy to real employees.
The findings arrive as AI becomes increasingly embedded in HR operations. Gartner reported that 95% of organizations had implemented AI in some capacity over the previous year, although only one in five had achieved significant or transformational value. Gartner also forecasts that AI could automate or perform up to half of today’s HR activities by 2030.
For HR teams, that makes the question less about whether AI should be used and more about where human judgment remains essential.
When polished language creates ambiguity
One of BrightHR’s clearest concerns involves seemingly harmless wording.
The AI-generated absence policy instructed employees to notify their manager “as soon as possible” if they could not attend work. O’Connor argues that such language leaves too much room for interpretation. An employee working an early shift, for example, may have a different understanding of “as soon as possible” from a manager expecting notification before the scheduled start time.
The policy also failed to make several practical points explicit, including who should be contacted when a manager is unavailable, whether text messages are acceptable and whether another employee can report an absence on someone’s behalf.
These details may sound administrative, but they are precisely the mechanisms that determine whether a policy can be applied consistently.
UK workplace guidance reinforces the importance of this level of specificity. Acas recommends that absence policies explain how unplanned absences should be reported, how unauthorized absence and lateness are handled, how trigger points work and what happens when employees return to work. It also says managers should be trained to apply the policy.
Trigger points are not automatic disciplinary decisions
The more consequential issue concerns absence thresholds.
The AI-generated policy stated that a third absence within a rolling 12-month period would result in a formal warning. That wording risks turning a review mechanism into an automatic disciplinary action.
Acas specifically distinguishes between an absence “trigger point” and disciplinary action. Trigger points can initiate an absence review, but employers should remain flexible and consider individual circumstances rather than automatically disciplining an employee once a threshold is reached.
That distinction matters because absence can involve circumstances requiring additional consideration, including disability, pregnancy-related sickness or other protected situations. A policy that presents a numerical threshold as an automatic consequence could therefore create problems if managers apply it without investigation or context.
The issue illustrates a broader weakness in using general-purpose generative AI for HR documentation: the system can satisfy the wording of a prompt without understanding the organization’s complete policy framework, management practices or employment-law obligations.
AI can draft the document. It cannot own the decision.
BrightHR is not arguing that employers should abandon AI for HR work. In fact, the company already uses AI across areas including recruitment and employee support. Its position is closer to a “human in the loop” model in which AI accelerates drafting and administrative work while qualified professionals remain responsible for review and decisions.
That approach is increasingly relevant as HR technology moves beyond chatbots and content generation into decision-support and automation.
The UK’s CIPD found that employees in 76% of organizations were using AI tools at work in autumn 2025, rising to 87% in the public sector. More than half of employers said employees were using free AI tools such as ChatGPT, Claude, Gemini or Microsoft Copilot.
In other words, HR departments are not introducing AI into an empty environment. Employees and managers are already bringing consumer and enterprise AI tools into workplace processes.
That raises the governance challenge.
HR leaders need to distinguish between low-risk applications—such as producing a first draft or summarizing existing material—and higher-risk applications involving employment decisions, disciplinary action, employee rights or legal compliance.
What enterprise HR teams should change
The practical lesson from BrightHR’s review is that an AI-generated policy should be treated as a working draft, not an approved corporate document.
HR teams adopting generative AI should test policies against existing procedures, define specific reporting channels and timeframes, establish escalation routes and ensure that any trigger-based process includes appropriate investigation and review.
They also need to check whether AI-generated language matches current legislation and the organization’s actual operating practices. BrightHR recommends training employees who use AI for HR work and maintaining human oversight, particularly where outputs could influence employment decisions.
This is becoming an important distinction in the HR technology market. Platforms from vendors such as Microsoft, Workday, SAP, Oracle and Salesforce are increasingly incorporating AI into enterprise workflows, while specialist HR platforms are adding AI assistants and content-generation capabilities.
The competitive question is therefore shifting from “Can AI generate HR content?” to “Can an HR technology system generate useful content while preserving governance, context and accountability?”
For enterprises, that may be the more important benchmark.
AI can produce a convincing policy in seconds. The harder task is ensuring that the policy still makes sense when a manager has to use it on a Monday morning, an employee challenges its application or an employment-law issue reaches escalation.
That is where human expertise remains difficult to automate.
Market Landscape
The HR technology market is moving toward AI-assisted workflows, but adoption is exposing a gap between automation and accountability.
Gartner says AI adoption across HR has accelerated sharply, with 95% of organizations implementing AI in some capacity during the previous year. Yet only 20% reported significant or transformational value, suggesting that simply deploying AI does not guarantee effective outcomes.
The policy-generation use case is particularly revealing because it sits between administrative automation and high-stakes decision-making. Drafting a document is relatively low risk; deciding how that document should govern sickness absence, disciplinary procedures or employee rights is not.
This creates an opportunity for HR software providers to differentiate through contextual intelligence, auditability, approval workflows, legal updates, role-based controls and human review rather than generative AI alone.
For enterprise HR teams, the emerging model is likely to be AI-assisted, human-governed HR. AI can reduce administrative workload, but organizations still need clearly defined accountability for employment decisions.
Top Insights
- BrightHR’s AI policy review found that polished HR documents can omit operational details, creating consistency and employment-law risks for managers and employees.
- Vague absence instructions around reporting deadlines, communication channels and escalation routes can undermine otherwise comprehensive policies when managers apply them in real workplace situations.
- AI-generated absence trigger points should initiate reviews rather than automatically determine disciplinary outcomes, particularly where individual employee circumstances require additional consideration.
- Gartner reports that AI is already widespread across HR, increasing the need for governance frameworks that combine automation with human oversight and organizational context.
- Enterprise HR teams should treat generative AI as a drafting and productivity tool while retaining human accountability for policy approval and employment decisions.
Join thousands of HR leaders who rely on HRTechEdge for the latest in workforce technology, AI-driven HR solutions, and strategic insights
