DEI was canceled. DAI arrived. How AI is redefining business inclusion in 2026

Hello everyone!

The word “diversity” is no longer neutral. By 2025, the Fortune 100 companies had eliminated it from 98% of their communications. Not because companies have stopped caring about the issue, but because the word has become a political minefield that can cost an organization contracts, investor relations, and, in the United States, even litigation.

The question for Latin American and global executives isn't, "Should we abandon inclusion?" It's more uncomfortable than that: "How do we continue doing the work without the terminology that defined it for two decades?"

There is a solution that works. And it involves AI.

The root of the problem

Employee Engagement (DEI) programs became political targets in the United States after the Trump administration's 2015 executive orders, which banned such programs at federal contractors and put pressure on private companies. The chain reaction was swift: Amazon eliminated employee resource groups and scaled back its Diverse Supplier Program. McDonald's suspended its representation goals. Meta dismantled its DEI team entirely.

The most revealing data comes from Harvard Law: companies not only stopped using “DEI”, but the terms “diversity”, “equity”, “inclusion” and “racial equity” also disappeared. A complete wipe.

The problem with that is obvious. Structural biases in hiring, promotion, and compensation don't disappear just because you change the program's name or stop mentioning it in the annual report.

What's really happening

The most strategic companies aren't abandoning the work. They're reframing it. The vocabulary gaining traction includes "belonging," "opportunity programs," "talent equity," "skills-first hiring," and "fair access." These aren't empty euphemisms—they're frameworks that connect inclusion directly to measurable business outcomes, making them far harder to attack politically.

And this is where AI fundamentally changes the equation.

Traditional EID programs relied on voluntary commitments, public declarations, and numerical targets. This made them visible and vulnerable. AI allows the same work to be done systematically—integrated into processes, without needing a cartel. It transforms bias into a technical problem that can be solved with data, and that is much harder to dismantle with a decree.

Three concrete uses of AI to support intersectionality

1. Talent selection without embedded bias

Language models can analyze job descriptions, identify wording that discourages certain profiles—women, first-generation college graduates, candidates from non-metropolitan areas—and suggest neutral alternatives. Beyond language, AI systems can detect patterns in historical hiring decisions where bias is statistical but not intentional.

2. Visibility of internal trajectories

AI can map who has access to high-profile projects, senior mentors, and development programs. If certain demographics are consistently excluded from these pathways, the system reveals this before it becomes a retention or reputational issue. It turns intuition into data.

3. Measurement without political exposure

Instead of publishing annual diversity reports that are now public targets, companies can internally measure progress toward equitable access and use that data for management decisions. The commitment remains, but the conversation stays within the boardroom.

CASE 1

Unilever — AI in recruitment, before it was trendy

Unilever has been using AI in its recruitment process for years. The most documented result: a 16% increase in the diversity of shortlisted candidates after implementing AI-powered interview software. But that number isn't the most interesting thing.

That's how they got to him.

Unilever used predictive models to retrospectively analyze its hiring decisions and identified statistical biases based on gender, ethnicity, and socioeconomic class that human interviewers had missed. They didn't call it a DEI program. They called it a quality issue in the selection process. That reframing changed the conversation internally and neutralized the usual resistance.

In 2024, they expanded a reverse mentoring program where junior employees work directly with senior leaders to discuss inclusion, digital fatigue, and intersectionality. The framework is “Equity for Impact”—not “diversity” or “DEI.” The practical result: Unilever maintains its core commitments while navigating the political climate with vocabulary that doesn’t jeopardize contracts or institutional relationships.

CASE 2

Walmart — From “Chief Diversity Officer” to “Chief Belonging Officer”

Walmart made the most documented rebranding in the industry: it renamed its top diversity executive Chief Belonging Officer and published its first “Annual Belonging, Diversity, Equity & Inclusion Report” in 2024 with that wording. Belonging first.

But the language change came with real changes.

In 2025, Walmart launched an AI interview coach, a tool that helps internal candidates—especially those in frontline roles—prepare for interviews and advance to positions of greater responsibility. The initiative stems from a concrete fact: 75% of Walmart's managers started as store associates. The AI coach is not a diversity program; it's an internal mobility program that, in practice, creates equitable access for the company's 2.1 million employees.

In addition, Walmart launched the Skills-First Workforce Initiative, a mapping of 11 million frontline jobs based on skills, not credentials. Eliminating college degree requirements where they are not needed has a huge intersectional impact—opening doors to groups historically excluded from the formal labor market—but the business case is strong enough to withstand any political climate.

What this means for leaders

The work of inclusion did not die with the word “diversity.” What died is the declarative version of that work: the annual reports full of commitments, the representation goals posted on LinkedIn, the campaign-named initiatives.

What works now is systemic. It's embedded in talent processes, not in parallel initiatives. It's measured in business metrics—quality of hires, retention, access to development—not in demographic counts.

Three questions worth asking today: Do our hiring processes have documented statistical biases, or are we just assuming they don't? Does investment in internal development reach all profiles equitably, or do we have blind spots? Does the vocabulary we use connect with business results, or does it still sound like a campaign?

One last thing

Intersectionality—the recognition that people have multiple, overlapping identities that affect their work experience—doesn't disappear just because companies stop using the word. First-generation college-educated women in technical or STEM roles continue to face different barriers than their male colleagues. Employees from non-metropolitan areas have different career paths. Transgender people still lack access to formal employment. LGBTQ people continue to experience harassment and discrimination in the workplace. That's real with or without a DEI program.

What's changing is how companies choose to intervene. By 2026, those that get it right won't do so with posters or public reports. They'll do it with data, AI, and internally designed processes.

That's not abandoning inclusion. It's making it harder to dismantle.

diego@dtconsultings.com

dtconsultings.com

Sources consulted:

Harvard Law School — DEI in Transition: 2025 Corporate Diversity Disclosure Trends

Walmart Corporate — 2024 Annual Belonging, Diversity, Equity & Inclusion Report

Walmart Corporate — AI Interview Coach Launch, 2025

Technology Magazine — How Unilever Uses AI & Digital Solutions

Diversity.com — Why Companies Are Reframing DEI Language in 2025

Gravity Research — 5 Findings That Defined DEI in 2025