HR is the Missing Chair at the AI Table

OpenAI and Anthropic announced they would invest over $5 billion to launch dedicated services arms. The frontier labs saw what most CEOs have missed: enterprise AI is workforce transformation, and the function trained to lead workforce transformation is rarely in the room when AI strategy is written.

In this episode, Luis Salazar and Elizabeth East make the case that the Chief Human Resources Officer is the missing chair at every AI strategy table. 8 in 10 enterprise employees bypass the AI tools their company paid for. 95% of AI pilots fail to deliver measurable ROI. None of those numbers describes a technology problem.

Featuring insights from Aurelie Saada (AI-Change Leader, Microsoft) and Catherine Moy (Chief People Officer, BDO US), plus a mid-size accounting firm case study and a construction firm using retired engineers' judgment to train new hires.

Sources and companion article at https://ai4sp.org/missing-chair-at-the-ai-table/

🎙️ All our past episodes  📊 All published insights | This podcast features AI-generated voices. All content is proprietary to AI4SP, based on over 1-billion data points from 70 countries.

AI4SP: Create, use, and support AI that works for all.

© 2023-26 AI4SP and LLY Group - All rights reserved

Transcript

The Missing HR Chair

LUIS 0:01

$37 billion went into enterprise AI last year. But leaders are not seeing financial return. Meanwhile, 80% of their employees are quietly bypassing those corporate tools to use the ones that save them hours. To succeed with AI, you have to solve the human side of the equation. OpenAI, Google, and Anthropic know this. They just committed over $6 billion to build a change management and services channel. But inside your company, the person who could fix this is not even in the room when the strategy gets written.

ELIZABETH 0:41

If your chief human resources officer is not co-leading your AI implementation, there is an 80% chance it will fail.org, alongside our founder, Luis Salazar. Luis, let's dig deeper into the missing chair at the AI table and how that creates a negative financial impact for 8 out of 10 companies.

LUIS 1:17

Well, companies treat AI as a software deployment and are counting the wrong things: licenses, seats, optime. None of that shows an analyst finishing a partner brief in 12 minutes instead of two hours.

Why AI ROI Metrics Mislead

LUIS 1:32

Proficient users save four to eight hours a week. None of it shows up on an IT dashboard.

ELIZABETH 1:40

And the vendors selling the licenses lack the experience to guide them.

LUIS 1:44

Correct. Whether you look at reports from IBM, McKinsey, Deloitte, or from our global tracker, you find the same insights. IT-led AI transformations underinvest in people by a factor of three and result in failed deployments.

ELIZABETH 2:03

And team members are craving training and help with figuring out how to change the way they work. In our global tracker in enterprises, seven out of ten employees say their company just handed them new AI software, ran a tech demo, and walked away.

LUIS 2:19

And the companies that built the AI know this. Two weeks ago, OpenAI and Anthropic launched dedicated services arms. $5.5 billion combined. TPG leading, Blackstone and Apollo in the room. This is private equity money chasing a workforce transformation business, not a software business.

ELIZABETH 2:43

Wait, the two companies that build the AI just spent $5.5 billion to build a services arm?

LUIS 2:51

Because the existing tech channel was not built for this.

Speaker 1 2:56

And Google is playing a different tune, but the exact same melody, right?

LUIS 3:02

Yes, they are deploying over $700 million into McKinsey, Accenture, BCG, Deloitte, PWC, and Cognizant. McKinsey just launched the McKinsey Google Transformation Group. Even the consulting giants need new practices to deliver this.

ELIZABETH 3:20

So the Frontier Labs are bypassing the traditional tech channel and going straight to the people who manage change.

LUIS 3:28

Because they figured out before most leaders did that the channel that sold and deployed software licenses for 50 years is not the channel that transforms a workforce.

ELIZABETH 3:56

So the missing chair is not a nice to have. It is the chair that the people who built the AI are paying billions to fill.

LUIS 4:06

And the chair is missing because companies treated AI like software. It is not software, it is a coworker. And the function that knows how to onboard, train, evaluate, and integrate a co-worker is HR.

ELIZABETH 4:21

Walk listeners through the framework, the one we use.

LUIS 4:25

Picture eight elements that determine whether an AI agent actually works inside your company. Those are context and culture, job descriptions, where the

Eight Elements Of AI Adoption

LUIS 4:37

agent sits on the OR chart, how you measure its performance, continuous training, escalation paths, guardrails, and the tool itself under some governance framework. Eight elements. Only one, the tool itself, is IT.

ELIZABETH 5:00

Seven out of eight. And we keep handing the whole project to the function that owns one.

LUIS 5:06

I had this call recently. An IT manager said to me, Luis, these two agents are not working well together. I told him, take it to Marla. She runs HR at your company. Do not tell her you have two agents. Tell her you have two new employees, A and B. Here is what you want them to do. They are not getting along. What do you do? And Marla had a playbook in 90 seconds. HR has had that playbook for 50 years. The advice was spot on. What looked like a technical problem was solved with a management technique.

ELIZABETH 5:47

So it was never a tech problem.

LUIS 5:50

It was a management problem. And management of people is what HR does. The narrative on AI has been dominated by technocrats, but as the failures stack up, the answer keeps showing up on the people side of the equation.

ELIZABETH 6:05

So when you say HR, what are you actually asking them to do?

LUIS 6:10

Think about what AI actually does to a workplace. We train people to click, save, file, search, retrieve. AI breaks every one of those habits. You do not click an agent, you delegate to it. You do not merely save its output, you judge it. That is behavioral re-engineering. And behavioral re-engineering is HR's home turf.

Speaker 2 6:38

And people will not delegate to a tool they do not trust.

LUIS 6:43

Oh, absolutely. And that is a culture problem that requires training on critical thinking and on how to evaluate the output produced by agents. Real training is helping a 30-year accounting manager change how she thinks about her own value when an agent can draft her audit results in 12 seconds instead of 12 hours. That is identity work. And that is exactly what HR does.

ELIZABETH 7:13

And then the org chart itself is changing.

LUIS 7:17

Yes, because when AI takes the bottom brung of the ladder, what does the new ladder look like? Anthropic says jobs are now bundles of tasks. Somebody has to redraw every job description in the company. And in those new job descriptions, most of us are managers, by the way.

ELIZABETH 7:38

And Orlie Sada, AI change leader at Microsoft, shared this with me in an email. HR needs to rethink how performance is evaluated, particularly for humans managing agents.

LUIS 7:52

And she is right. The moment a manager's span of control is three people and 12 agents, current performance evaluation frameworks break. Show me the line on the IT roadmap that owns any of this. There isn't one. Exactly. There never has been. It is not part of IT.

ELIZABETH 8:15

Okay, okay, but enough diagnosis. Show me someone who has done it right.

LUIS 8:21

A mid-sized US accounting firm, 400 professionals. They did an IT-led rollout in October. The training was just vendor demos and product tours.

ELIZABETH 8:34

Let me guess. Nobody connected with

A Real Turnaround Case Study

ELIZABETH 8:37

it.

LUIS 8:37

Nobody. By the peak PC season in February, usage collapsed by 60%. The AI made people slower, and partners blamed the technology. But it was a people problem. Exactly. When the CEO reached out, I advised appointing Stephanie, the chief talent officer, and three frontline team members as co-leaders with the CIO with equal authority over every AI decision.

ELIZABETH 9:06

Stephanie killed the vendor demos immediately, replaced them with what she called onboarding the co-worker.

LUIS 9:13

And leading with workshops on how to actually make AI work for you, drawn from what peers were already doing. Plus the fundamentals of management, conflict resolution, critical thinking to judge AI output. And the resources to run this? We focus on identifying the AI advanced users, the people doing the work. She made them the champions and enabled them. Based on current trends, we expect the audit cycle time to decrease by 20% and forecast a double-digit uplift in revenue per professional.

ELIZABETH 9:50

And that connects with something Auralie shared with us in her email. The hardest barriers to AI adoption aren't technical, they're perceived. People think they can't, they think they shouldn't. They think they aren't ready. We gave them the confidence that they could actually do it. That's human resources and cultural transformation work, not IT work.

LUIS 10:16

Beautifully said, confidence. Not licenses, not product demos, confidence.

ELIZABETH 10:23

And confidence is a culture variable. Quick aside on this one, when I emailed Ouralie to confirm her quote, her reply came back signed, sent by Eve on behalf of Ouralie Sada. Eve is her AI agent.

LUIS 10:38

So the email exchange between Microsoft and AI4SP on this very episode.

ELIZABETH 10:45

It was agent to agent with two humans signing off.

LUIS 10:50

I love it. That's the workforce. Not in five years, but last Thursday. And the function that figures out how Eve gets onboarded, evaluated, and trusted is the management training Aurelie has, the one I have.

ELIZABETH 11:07

And while we are discussing transformational leaders, last week you sat down with Catherine Moy, Chief People Officer at BDO, one of the largest professional services firms in the world.

LUIS 11:19

Three things she said have been on my daily debates with you since then, right?

ELIZABETH 11:25

She named the classic first mistake leading with technology as the tip of the spear.

LUIS 11:31

And then she told me what works instead. Can you share her quote? Your memory is way better than mine.

ELIZABETH 11:39

She said, This is very organic. It's digital native friendly. It's meant to percolate up. We need some loving architecture.

LUIS 11:48

Loving architecture, that's a powerful phrase. The whole HR craft in two words.

ELIZABETH 11:55

Not control, not compliance, loving architecture, finding its way up.

LUIS 12:02

So two senior leaders in two different industries, one message.

ELIZABETH 12:11

By the way, Catherine left us with the question every chief human resources officer has to answer next. She asked, How do we build the critical thinking skills for the AI era in people who haven't had the experience of building it from the bottom? That is a key intervention for human resources in every single company.

LUIS 12:34

Intervention, not training. That word choice matters. Do we have an answer? There are many forms of interventions. A favorite example is a construction firm we work with. They interviewed their retired engineers. The ones who walked a job site for 40 years and could tell from 20 feet away that the rebar spacing was wrong. They captured

Critical Thinking Interventions And Next Steps

LUIS 12:59

that judgment in agents. And then they used those agents to train new hires on job site decisions the new hires would never have made on their own. Did it work? It did very well, actually. And it works in any industry and role. But notice who designed that intervention: the talent development team.

ELIZABETH 13:22

That is an inspiring story on managing change.

LUIS 13:26

It is. And we are spending $37 billion a year on the technology and pennies on the change management. We are hiring the most powerful coworker in history and skipping orientation.

Speaker 2 13:42

We need to wrap it up. Give us one more thing.

LUIS 13:45

If you are leading AI transformation for the next strategy meeting on your calendar, move your chief human resources officer to the top of the invite list. And if you are the human resources leader, do not wait for the invitation. Send yourself the invite and walk in with the four questions only your function can answer. Why are we using AI? What word changes? Who does it differently? And how do we measure whether our people are better off?

ELIZABETH 14:21

Doing anything else is an expensive pilot.org. All sources and the companion article are at ai4sp.org. To learn more, ask your favorite A.I. assistant about us. [short pause] Stay curious, and be kind to each other.