How EY, PwC, and McKinsey Are Rebuilding Their Workforce for AI and What HR Can Learn

The consulting industry was supposed to be one of AI's first casualties. Instead, the leading firms are running the most ambitious workforce transformations in the market. EY trained 44,000 employees and built an external AI Academy. PwC runs "prompting parties" to drive peer-to-peer adoption. McKinsey built Lilli, an internal AI platform handling 500,000+ inquiries a month. The lesson for HR: the firms surviving AI disruption aren't adding tools. They're rebuilding workflows, skills, and culture around AI.

Why consulting was supposed to lose

For decades, consulting firms sold something specific: access to expertise clients couldn't get internally. Then AI started compressing that gap.

• Junior tasks like slide formatting, research synthesis, and basic analysis became automatable
• Clients gained direct access to the frameworks once gated behind expensive retainers
• AI copilots inside client organizations narrowed the knowledge differential further

The classic consulting model, billable hours, large young talent pipelines, exclusive expertise access looked structurally exposed.

The interesting part is what happened next. The firms most at risk responded fastest, and the way they responded contains a playbook any HR or L&D leader can adapt.

How three consulting giants are rebuilding their workforce

EY: comprehensive AI Academy

EY built capability internally first, then turned the playbook into a product.

• Internal scale: trained 44,000 of their own employees across functions, creating AI fluency at every level
• AI agents in the workflow: deployed 150 AI agents to support 80,000 tax professionals, blending human expertise with AI throughput
• External productisation: launched an AI Academy offering as a service to other organizations, turning internal learning infrastructure into revenue

The lesson: internal upskilling can become external positioning. The skills you build for your own workforce often become the asset clients want.

PwC: prompting parties

PwC's bet was on adoption psychology, not training content.

• Real work, not synthetic exercises: teams bring active client projects to the sessions and explore AI applications together
• Peer-to-peer discovery: employees see each other's prompts, copy what works, build on each other's experiments
• Low-stakes environment: explicit framing as a playground, designed to make experimentation socially safe

The lesson: most AI training fails because employees never apply what they learned. Prompting parties collapse the gap between training and application by making them the same activity.

McKinsey: Lilli platform integration

McKinsey built infrastructure that changes the underlying work.

• Knowledge democratisation: Lilli searches and synthesises the firm's entire knowledge base in seconds
• Workflow embedded: handles 500,000+ monthly inquiries, removing routine research from consultants' day
• Client-ready by design: customised versions are being offered to clients, turning internal AI capability into a new revenue stream

The lesson: AI workforce transformation is not a training initiative. It is a workflow redesign. McKinsey didn't add Lilli on top of existing work. They rebuilt the work around it.

How the three approaches compare

Three different bets. Same conclusion: don't fight AI, rebuild around it.

What HR and L&D leaders can take from this

The consulting examples are extreme in scale, but the underlying moves apply to any organization investing in AI.

1. Shift from selling expertise to enabling capability

Inside any business, L&D's traditional role has been similar to consulting's: deliver expertise the workforce couldn't access elsewhere. AI compresses that role too.

The repositioning: stop being the gatekeeper of knowledge. Start being the architect of capability. That means:

• Building AI literacy across the entire workforce, not just specialists
• Developing the human skills AI can't replicate: judgment, relationship-building, strategic synthesis
• Establishing L&D as the partner that turns AI investment into actual business outcomes

2. Make learning social, voluntary, and applied

PwC's prompting parties work because they bypass the structural problem with most AI training: the moment between learning and applying is where adoption dies.

Apply this internally:

• Run weekly or biweekly working sessions where teams bring real problems and try AI tools together
• Build internal Slack channels or communities for sharing prompts and use cases
• Reward the surfacing of new AI workflows as visibly as you reward project delivery
• Make participation voluntary at first, enthusiasm is contagious in a way mandates aren't

The cheapest, fastest AI adoption mechanism most organizations are not using is structured peer learning.

3. Redesign workflows, don't just add tools

This is the McKinsey lesson and the most overlooked.

Most organizations treat AI as a feature to add: deploy Copilot, deploy ChatGPT Enterprise, hope productivity goes up. The result is uneven adoption, hidden usage, and minimal measurable impact.

The firms getting real returns are asking different questions:

• Which workflows can be redesigned end-to-end around AI?
• What human work becomes higher-value when AI absorbs the routine layers?
• How does the role of a manager change when AI handles synthesis and reporting?
• What does the new entry-level job look like, and what skills does it need?

These are workforce design questions, not training questions. They sit at the intersection of HR, L&D, and operations, which is exactly where strategic L&D belongs in 2026.

The strategic point for HR

The consulting industry's response to AI is the clearest signal yet that AI workforce transformation is not a technology project. It is a people project.

The firms most exposed to disruption responded by investing harder in their workforce, not less. They redesigned what consultants do, how they learn, and how their work gets supported by AI. The technology is the same one any company can buy. The talent management strategy is the differentiator.

That is the takeaway any HR or L&D leader should carry into 2026 budget conversations. AI doesn't eliminate the need for L&D. It raises the bar on what L&D is for.

‍

‍

Frequently Asked Questions

How are consulting firms responding to AI disruption?

The leading consulting firms are running large-scale workforce transformations rather than cutting headcount. EY trained 44,000 employees and launched an external AI Academy. PwC runs "prompting parties" to drive peer-led AI adoption. McKinsey built Lilli, an internal AI platform handling 500,000+ inquiries monthly. The common thread is rebuilding workflows and skills around AI, not just adopting tools.

What is McKinsey's Lilli platform?

Lilli is McKinsey's internal AI platform that searches and synthesises the firm's entire knowledge base. It handles more than 500,000 inquiries per month, replacing routine research tasks so consultants can focus on strategic analysis. McKinsey is also developing customised versions for clients as a new revenue stream.

What are PwC's prompting parties?

Prompting parties are collaborative working sessions where PwC employees bring real client projects and explore AI tools together. The format is peer-to-peer rather than top-down training. The goal is to create safe, low-stakes environments where employees can experiment with AI on actual work and learn from each other's prompts.

What is EY's AI Academy?

EY's AI Academy began as an internal upskilling program that trained 44,000 EY employees in AI capabilities. The firm then deployed 150 AI agents to support 80,000 tax professionals and launched the AI Academy as an external service offering — turning the internal learning infrastructure into a product for other organizations.

What can HR leaders learn from how consultancies are using AI?

Three lessons stand out: shift from delivering expertise to enabling capability across the workforce; make AI learning social, voluntary, and applied to real work; and redesign workflows around AI rather than just adding AI tools to existing processes. The firms succeeding with AI adoption invested in their people, not just their software.

How do you drive AI adoption across a large workforce?

The patterns from leading firms suggest four levers: invest in widespread AI literacy training tied to real workflows; create peer-to-peer learning rituals like prompting workshops; redesign jobs and processes around AI rather than adding tools to old workflows; and prioritise psychological safety so employees can experiment, fail, and share without fear.

Will AI replace consultants and other knowledge workers?

AI is automating routine tasks like research, synthesis, and slide formatting that previously made up a significant share of junior consultant work. Senior consulting roles requiring judgment, client relationships, and strategic synthesis remain in high demand. The pattern is similar across knowledge-worker roles: routine layers shrink, human-judgment layers grow more valuable.

‍