DEI and Inclusion Consulting in India

The pattern

Most inclusion work sits in one of two places. The first is awareness: a calendar of events, a statement of intent. It does what awareness is designed to do, and was never scoped to change who gets hired, developed or heard. The second is a framework built for a different structure.

Representation at entry level looks reasonable, sometimes better than your sector. It thins at first-line management, thins again at the layer below the executive committee, and at the top it has not moved. Meanwhile the programme is renewed annually and the shape of the pyramid holds.

01Healthy diversity at intake that does not survive the first promotion gate.
02Exit interviews from under-represented groups that cite fairness of progression, ahead of pay or role.
03Sponsorship concentrated in a handful of senior leaders who each back people who resemble them.
04A screening tool the vendor says is bias-tested, with no statement of what it weights for the hiring chain to check.

THE GLOBAL PICTURE

86%

of business leaders say embedding diversity, equity and inclusion into everyday ways of working while measuring outcomes is important or very important, while only 25 percent feel very ready to address it, across 10,000 business and HR leaders in 105 countries.

Deloitte, 2023 Global Human Capital Trends

31%

of respondents still feel diversity is a barrier to employee progression at their own organisation, across roughly 6,000 responses in more than 52 countries and more than 26 industries, across roughly 6,000 responses in more than 52 countries and more than 26 industries.

PwC, Global Diversity, Equity and Inclusion Benchmarking Survey, July 2022

16%

of organisations have diversity, equity and inclusion goals that influence performance evaluation and compensation outcomes for leaders, and 26 percent report none of the accountability practices surveyed at all, in the same PwC sample.

PwC, Global Diversity, Equity and Inclusion Benchmarking Survey, July 2022

All three figures describe what organisations state and what they have installed, not what changed for anyone inside them. Deloitte and PwC both survey leaders and programme owners, so read the set as a reasonable measure of the distance between commitment and mechanism and a poor measure of lived experience. The progression figure is a perception reported by respondents, not an audit of promotion outcomes, and it establishes no causal claim about why any funnel narrows. None of them can tell you where yours narrows, at which gate, or for which cohorts. That is what three years of your own data is for.

The decision points that set your representation

Representation is the arithmetic result of these decisions, repeated across a year. Each is a designed process with an owner, which makes it available to change.

Why it persists

Awareness events and a borrowed framework both work on what people know. Representation is set by decisions taken elsewhere. Who gets hired depends on how the job description is written and who sits on the interview panel. Who gets promoted depends on whether readiness is judged against a defined standard or against proximity to the people judging.

Each of those is a designed process with a named owner. Each is also a moment where judgement is exercised under time pressure, when pattern-matching substitutes for assessment. Awareness training works on the person making the judgement. Changing what the moment demands as evidence works on the judgement itself, and holds when the person changes.

Artificial Intelligence (AI) in hiring is a decision point of the same kind, differing in one respect. An automated screen sits inside the applicant tracking system, and its weighting is usually documented nowhere in the hiring chain. It was procured rather than designed. Its assurance is a vendor claim written for a procurement form, so algorithmic bias becomes a question with no owner. Three questions, as for the others: who owns it, what does it weight, and can an outcome be explained to the candidate it affected, in language that candidate would accept. In hybrid and remote organisations, proximity bias works the same way: who is visible has changed.

Progression runs through a forum you already hold, on a cadence you already control. Accountability inside it is usually personal rather than structural: a sponsor's conviction, an HR leader's persistence, a champion network's energy. When the sponsor moves, nobody is obliged to keep the item on the agenda, and the calibration forum runs as it ran before.

Bias enters at designed decision points, each with a named owner and a documented process. A process can be rewritten and audited in a way that a belief cannot, and it holds when the owner changes.

What the work examines

POLIP starts from the identities people actually carry into work. In multi-site Indian businesses, language and region often shape access to senior forums more than any category a global template would flag.

Three years of hiring, promotion, pay and exit data are analysed by cohort: where the funnel narrows, at which gate, for whom. That produces a map of decision points with the evidence attached, and which gates carry enough weight to redesign. Where the analysis shows no pattern at a gate, that is itself the finding.

A panel member, a calibrating manager and a sponsor are asked the same things in the same order, which makes their answers comparable and the finding defensible when it is challenged. Where a pattern traces back to a named sponsor, it is reported to that sponsor's own forum rather than routed around them. Which gates come into scope is agreed with you in scoping.

Cohort and funnel diagnostic

Hiring, progression, pay and exit data analysed by cohort, to locate the gates where the funnel narrows and size each one.

Decision-point redesign

Job requirements, panel composition, calibration protocol, promotion criteria and any automated screening tool examined so the moment itself demands evidence, heaviest gates first.

Sponsorship architecture

Stretch mandates, visible projects and senior advocacy examined as a management obligation with named owners, not informal affinity.

Accountability that survives transition

Representation and progression placed in the forums your leaders already answer to, on the same cadence as commercial performance.

Where this shows up in the business

Day to day
Decisions acquire a record
Within weeks a hiring manager notices that panels are composed rather than assembled, and that a calibration claim without evidence gets asked for evidence.
Over a year
The numbers your leadership team already tracks
The promotion ratio at your own first gate, internal fill rate for senior roles, regrettable attrition in the populations you are trying to retain. A board can ask about these without a separate reporting pack.
Over several years
A pipeline that stops narrowing at the same gate
A funnel that loses the same cohorts at the same point every year is a capacity constraint on growth, priced quietly as a recruitment cost. Representation holds through a change of leadership because the mechanism holds rather than the sponsor.

POLIP publishes no client results and promises no figure in advance. When a board or a regulator asks what your organisation does about equity, a documented redesign of hiring, calibration and promotion is an answer with owners and a record.

Common questions

How is this different from the DEI work we have already run?

Awareness work, meaning training, networks, a statement of intent and a calendar, is built to change what people know, and it does that. Redesigning the hiring, calibration and promotion decisions themselves is separate design work.

POLIP works on the decisions. The deliverable is a redesigned job requirement, a composed panel, a calibration protocol and a promotion standard, each with a named owner and a record. If the work already in place has changed those decisions, the honest answer is that you do not need this, and POLIP will say so.

Which identities does POLIP work with in an Indian context?

The ones people actually carry into your workplace, a wider set than a global framework usually accounts for: gender, generation, tenure, socioeconomic and educational background, disability, language, and regional identity.

Scope is set from your own data and your own structure, in the scoping conversation, rather than from a standard list.

Do you set representation targets?

Targets without mechanism tend to produce either a hiring push that reverses or a quiet reclassification of existing headcount. POLIP examines the mechanism first, then discusses targets the redesigned process could plausibly carry, at each level rather than in aggregate.

What any target should be, and whether targets are the right instrument for your structure at all, is a scoping question. The test POLIP applies to its own work is whether representation holds at every level, including the two layers below your executive committee, and whether it survives a change of leadership.

Does AI in hiring reduce bias, or does it hide it?

Either, and the difference is whether anyone can account for the outcome. An automated screen is procured, so its weighting sits with the vendor, not an owner in the hiring chain, which is how algorithmic bias becomes invisible rather than absent.

POLIP puts any automated screen or ranking into the same decision-point inventory as the job description, the panel and the calibration forum. POLIP does not audit AI models and does not advise on legal obligations; the work stays on who owns it, what it weights, and whether an outcome can be explained.

Diversity, Equity and Inclusion Consulting