Chief Human Resources Officer / India
A promise needed different working conditions
- People-function colleagues
- 26
- Four managers; ₹5.3 crore annual function expense
- Company employees
- 1,180
- March 2026; Bhubaneswar, Mysuru and distributed Indian locations
- People experience
- 26 years
- Personnel coordination through executive responsibility
Decisions & consequences
01Replacing live client work in a candidate exercise
In 2022, a review of recruitment for specialist localisation roles found that several final-stage exercises used text from live client assignments. Candidates were asked to return a completed sample without payment, and the expected time was not consistently stated. Managers valued realistic material and wanted evidence beyond an interview. The difficulty was that the exercise could take several hours and the output sometimes resembled work the company would otherwise pay an employee or supplier to perform.
I asked recruitment and delivery to examine the task being assessed and the subsequent use of each sample. We stopped the live-material route while that review was completed. The alternatives were to shorten the exercise, use controlled non-client material or agree a paid final-stage assignment with clear terms. We chose a combination: a bounded routine assessment using purpose-made material, and payment where a longer exercise was necessary for a defined role. Qualified legal colleagues reviewed permissions and the relevant candidate terms. I did not ask managers to abandon practical evidence; I asked them to separate assessment from production.
The revised process cost ₹12 lakh in its first year, including material preparation, assessor time and candidate payments. Hiring teams also lost the convenience of taking the next available client task as a sample. We tracked expected duration and asked candidates whether the exercise matched that description. Two exercises were shortened after feedback showed that their instructions required more preparation than the assessment intended. The changes did not establish that every candidate experienced the process identically or that every hiring decision became correct.
My responsibility was to fund a credible alternative and make its use part of recruitment practice. Delivery managers still judged the relevant capability. The lesson was that realism is not sufficient justification for an assessment method. A company should be able to explain the burden it places on a candidate, the evidence it needs and what it will do with the work submitted, including where the answer requires payment rather than another statement about opportunity.
Revisiting ratings supported mainly by proximity
During the 2024 appraisal cycle, people partners noticed repeated comments about employees being visible, immediately available or active in office discussions. Those observations could matter for some responsibilities, but several files did not connect them with agreed work outcomes. We reviewed 146 appraisals across selected distributed and office-based teams. Thirty-one relied mainly on proximity or availability language and lacked enough specific evidence to explain the proposed rating.
Some managers feared the review would presume that every lower rating for a remote colleague was unfair. I made clear that the exercise was about the basis of the decision, not a predetermined distribution of scores. We asked managers for task and contribution examples, gave employees a route to provide relevant context and used a separate reviewer where the original account remained unclear. We did not infer performance from login hours or replace managerial judgement with a new automated score. People colleagues facilitated the review while accountable managers explained the final decision.
Eighteen ratings changed. Thirteen retained their outcome after the evidence was made explicit. Associated pay and progression adjustments for the changed decisions cost ₹19 lakh in the reviewed cycle, with another ₹8 lakh spent on review and manager support. Those figures describe the defined exercise, not the cost of every appraisal issue in the company. The initial sample also did not establish a company-wide causal estimate of bias. We reported its limits alongside the corrections and checked the next cycle for the same unsupported language.
I had previously supported a form that asked managers to summarise leadership presence without explaining what evidence belonged under that heading. We changed it. The case reminded me that people teams can create the ambiguity they later criticise managers for using. A fairer review required better questions, an opportunity to add evidence and a willingness to preserve an unchanged outcome when its basis was sound. The objective was an explainable decision, not a more reassuring average rating.
Funding the workload behind a return-to-work promise
In 2025, twenty-eight employees were preparing to return after extended personal leave. Their circumstances varied and were handled privately. Our policy offered a conversation about transition support, but delivery managers had no common way to describe the work that would be covered while an agreed adjustment operated. Several plans were encouraging in tone yet depended on colleagues absorbing duties informally or the returning employee resuming a full workload earlier than discussed.
I worked with people partners and delivery heads to define a twelve-week arrangement for this group, with individual terms rather than one assumption about everyone's needs. Each plan identified the responsible manager, initial duties, temporary cover, review dates and a route to revise the arrangement. Qualified specialists addressed employment and health matters where relevant. We did not require employees to explain private circumstances to their whole team. Managers communicated the allocation of work without making personal disclosure a condition of support.
The company approved ₹34 lakh for temporary cover and manager capacity. Twenty-four employees completed the planned transition within twelve weeks. Four needed revised arrangements following review; we did not remove them from the account to report a perfect completion rate. The programme was not evidence that the same timetable would suit every future return. It showed that a defined group could receive a clearer and better-funded response than the informal arrangements available before it.
The trade-off was visible capacity cost and some delayed discretionary work. Delivery leaders accepted those consequences before the plans were described to employees as agreed. I also asked the people team to check whether the person and manager understood the same commitment at the start and at each review. A return arrangement can change for good reasons, but the change should be discussed and owned. Support becomes less credible when its practical conditions are left for the employee to negotiate after returning to work.