By: Abhishek Sen
In recent times, Chief Human Resources Officers (CHROs) have seen AI draft job descriptions, answer employee queries, summarise policies and support performance feedback. The focus for many has now shifted from what AI can do to the results it can deliver. AI has moved beyond HR demonstrations, yet many organisations still struggle to generate measurable workforce and business outcomes.
Vendors continue to introduce new capabilities as HR functions test use cases and leaders expect results. Yet weak data, outdated processes, limited skills and unclear accountability often prevent pilots from delivering value.
Recent research highlights the gap between adoption and advanced use. The EY Work Reimagined Survey found that 88% of employees use AI, but only 5% apply it in advanced ways, leaving up to 40% of potential productivity gains unrealised.
Why value remains difficult to capture
The main barriers to AI in HR lie beyond the technology. Fragmented skills records, outdated job structures and employee data across disconnected systems reduce output quality. Workflows built around manual approvals, repeated handoffs and workarounds can make inefficient processes faster without improving outcomes. AI use in hiring, compensation and performance decisions places aspects like fairness, transparency, privacy and regulatory compliance under HR accountability.
HR teams understand people and culture but not everyone may have the technical knowledge to evaluate solutions, challenge vendor claims and guide adoption. This is where employees need clarity on where to use AI and where to fall back on human judgement.
On one hand, technology cannot address weak data, dated processes or unclear accountability. On the other hand, building every capability internally can consume time and divert HR from core priorities. Together, this raises a practical business question around what HR should own versus where outside support can add value.
Build versus buy
The key is to own continuous capabilities and seek support for episodic, specialised work.
Capabilities to own within HR
Build AI fluency through weekly Human Resources Business Partner (HRBP) practice, shared prompts and output reviews.
Base use-case prioritisation on workforce needs, organisational context and data gaps.
Retain responsible prompting, content development and AI use, with people accountable for outputs and decisions.
Own vendor relationships, roadmap planning and renewals, using support for selection and negotiation.
Where external support adds value
Design role-based AI academies with company cases, labs, then transfer delivery to HR.
Redesign job and skills architectures, followed by internal reviews.
Assess vendors, test claims and support negotiations using implementation experience.
Design governance covering bias reviews, explainability, DPDP cross-border data compliance and works council requirements.
Remove outdated approvals and handoffs before automation.
Integrate AI agents across HRIS, ATS, payroll and ITSM.
From connected workflows to measurable outcomes
Assessing vendors under real operating conditions rather than polished demonstrations, is critical. Before selecting an AI solution, ask whether it works with existing job architecture, role titles and incomplete skills data. What technology powers it? How does it connect with existing workflows and the ATS? What information can it read and write back? Where does the vendor process and store company data? Does company data train its models? Can the organisation move its data and change providers without high costs or restrictive contracts?
The answers reveal product quality, workflow fit, data governance and long-term flexibility before commitment. Testing products with company data shows the gap between a controlled demonstration and daily performance. Independent assessment helps CHROs compare products, test vendor claims and identify long-term costs.
After implementation, lasting value depends on workflow integration, clear ownership, adoption and accountability. Workforce planning, hiring and performance management continue to need human judgement and defined accountability.
In the first 6 months, track active usage, accuracy, query deflection, draft acceptance and time saved.
From 6 to 18 months, measure hiring, onboarding and query-resolution times, transaction costs, employee experience and time moved from administrative to advisory work.
After 18 months, assess HR costs, quality of hire, internal mobility, retention, workforce productivity and employee trust.
The leadership imperative
CHROs face a broader decision than selecting a product. They need to determine what HR should own, what technology to buy and where outside support can close capability gaps. Poor choices can create disconnected investments, low adoption and pilots that never enter daily operations.
Measurable value depends on prepared data, redesigned processes, informed teams, defined governance and agreed measures of success. A focused build-versus-buy approach keeps lasting capabilities within HR and directs outside support toward work where wider implementation experience adds value. Human judgement and accountability remain central.
