I would help a board examine whether the evidence in a proposal is suitable for the decision it is being asked to make. A prediction may use knowledge that would not exist at the relevant time. A growth rate may compare different populations. A complete-looking chart may conceal absent history. I can follow those questions with management and ask for an explanation that a director can assess without rebuilding the analysis. My contribution would be useful where a company is extending analytical tools into consequential operating decisions and needs to understand their limits. I would also ask what a report allows its audience to infer about people, rather than assume that removing names resolves every concern. This is a particular perspective, not a substitute for the board's financial, legal or sector expertise. I have no prior directorship and would need to learn the company's wider model. The value would be in improving collective judgement about evidence and accountability while preserving management's responsibility for the work.
Judgement in practice
01
Letting a lower score stop a premature release
The original prototype score looked persuasive until we examined the timing of its inputs. Removing later-event fields and using a later-period evaluation produced a lower result, and we withheld release. I kept both changes visible rather than attributing the entire difference to one cause. At board level, I would ask whether a model has been evaluated under conditions that resemble its proposed use and who has authority to approve that use. I would not select the algorithm or set credit policy from the boardroom. The useful challenge is to establish whether management's claim answers the actual decision, whether the limits are understood and whether the responsible reviewers can disagree without being treated as obstacles to delivery. A technically impressive score should begin an examination of evidence rather than end it, particularly when the planned use affects customers or other consequential relationships.
02
Giving up detail that the general audience did not need
The village-cluster dashboard had no names, but narrow combinations of fields made some groups small enough for local readers to infer individuals. I supported broader groupings and a controlled route for finer analysis after review with the responsible owners. That reduced convenience and made some enquiries slower. In a board role, I would ask how management connects the audience and purpose of a report with the detail it distributes. I would want to know who judges the permitted use and whether the review considers combinations of outputs, not only each screen in isolation. I would avoid claiming that a fixed minimum group size settles every privacy question. The board's role is to seek an accountable process and credible evidence of its operation, while management and qualified specialists determine the controls appropriate to the company's particular circumstances.
My first year on a board
I would start by learning how the company makes money, serves customers and delegates consequential decisions. With the chair's agreement, I would examine a few representative board papers and ask how their main analytical claims were produced and reviewed. I would meet finance, operations, risk and technology leaders to understand where definitions and responsibilities differ for legitimate reasons. I would also read the committee terms before suggesting additional reporting, so my interests did not create a parallel data-governance exercise disconnected from the board's needs. In the first annual cycle I would focus on whether directors can identify the population, period and limitations of evidence used in major decisions. I would seek induction on matters outside my experience and feedback from the chair about the usefulness of my questions. I would need to resist the executive habit of offering to repair the analysis myself; the board should be able to rely on a management process that works without a director becoming its additional analyst.
Conflicts and their management
The bank's customers, suppliers and service relationships could connect with a prospective company even where the sectors appear different. I would disclose relevant interests and obtain employer consent before proceeding. I would not bring customer-level examples, confidential models or unpublished bank measures into another board's discussion. Nor would I offer access to my team or the bank's records as an informal benefit of appointment. Independence and the applicable directorship and committee limits would require assessment for the proposed company. If a relationship would repeatedly prevent me from participating in its central decisions, I would question the suitability of the seat rather than rely on routine recusal. The eighteen-day allowance includes reading and travel and must be checked against the actual calendar and likely additional demands. I would be candid about the limits imposed by my executive role before accepting responsibilities I could not reliably discharge.
My independence statement
Any proposed company appointment would require review of my employment, interests and relevant relationships. I would make the disclosures needed for assessment under section 149(6) and the other requirements applicable to that company, including SEBI LODR where relevant. This statement concerns a prospective company seat and does not assume that my current cooperative employer has the same governance framework.
Board and committee exposure
Presented to the Board or a committee
Varneli Cooperative Bank
Invited presentations on data quality and analytical use
2021–present
Explained management-reporting limitations, data ownership and proposed analytical controls; no director or committee-member position.