Angela Beatty, former Chief Leadership and Human Resources Officer at Accenture, on AI-enabled workforce transformation

As AI moves from experiment to operating system, Angela Beatty argues that CHROs must treat it as a new way of working, not a tool to install. Drawing on her experience leading large-scale transformations at Accenture, she discusses how generative AI can reinvent performance feedback, learning, talent acquisition, and HR service delivery. She also outlines what CEOs and boards most need to grasp about AI transformations of work and the workforce.

Angela Beatty is an innovative HR leader and advisor working at the intersection of business, talent, and technology. Most recently, she served as chief leadership and human resources officer at Accenture, responsible for 750,000+ people in 120+ countries, where she led talent transformations including large-scale AI talent rotation and reimagined performance, rewards, and HR operations. She is a member of the Women’s Leadership Board at Harvard’s Kennedy School, HR50, the CHRO Association, and the Center for Executive Succession.

To start, since you’ve led AI deployments at scale, you have a clear vantage point on where the buzz on AI matches reality and where it doesn’t. This is sought after experience, and I know boards and CEOs are coming to you with real-world questions. What are they asking you, and which of those have clear, practical answers versus genuine uncertainty?

Boards and CEOs tend to ask some version of the same four questions: One, where is the value and how fast can we see it? Two, what are the risks and the controls we should put in place? Three, do we have the data, skills, and operating model to scale? Four, what happens to our workforce?

On the value and evidence side, we already have solid, practical answers. Repeatable wins are common. For example, in new business proposal drafting, we know the time can be cut down by around 85 percent at scale. I’ve seen 35 to 50 percent productivity gains with the use of co-pilots. Organizations that modernize work and talent together and embed AI in the core are able to scale faster and perform better. Upskilling and skills rotation through AI-fueled skills engines, leveraging a common skills taxonomy for the enterprise and AI-inferred, human-confirmed proficiency and learning programs, can be the foundation for educating, enabling, and embedding AI and leading to fast transformations at scale. On governance and responsible AI, organizations that have really stood this up, doing things like quarterly AI talent audits, putting responsible AI controls in place, running bias checks, and having humans in the loop for consequential decisions, are showing that these practices are both proven and imperative.

The pushback usually shows up around constraints. Some organizations will say, we are a smaller PE-backed company, we do not have all the resources to invest in an AI and agentic overhaul, and we need to see results quickly. To that I say, start with three to five workflows, adopt a single skills language, stand up lean governance, and you are probably going to get material impact within a few months. A second constraint is data readiness. Data is essential to this, and if it is not in good shape, that becomes the priority. You can start with co-pilots and local data domains and mature that digital core over time. In some regions, particularly in parts of Europe, regulatory, union, and works councils complexity also plays a role. That makes it even more important to keep humans involved and in charge of decisions, make sure agent actions are auditable, and use it as an opportunity to engage unions and works councils to co-design the change.

The genuine unknowns relate to the pace of agent autonomy: in regulated settings, how liability frameworks will evolve, and how quickly labor pyramids will reshape by sector. Organizations can manage that uncertainty by putting in staged guardrails, running experiments and testing use cases, having scenario-based workforce plans, and again keeping humans in charge, building trust, and designing for employee experience from the start. That also has a big impact on improving adoption and the impact you ultimately realize.

Let’s dive into a concrete example of AI in action, to give us a practical case to think with. You’ve led the transformation of the performance feedback process at scale and I’m sure are advising others on this and similar cases. Can you describe what using AI in performance feedback at scale can look like, and, based on your experiences, which other core HR areas organizations should review as potentially most ready for AI-driven transformation?

Transforming performance feedback through AI is a great example of how AI can make a time-consuming, sometimes burdensome process faster, higher quality, and a better experience for everyone.

What can this look like in practice? First, set up an AI-powered feedback coach to help people write effective comments in minutes. The coach prompts managers to focus on specific skills, collaboration, and opportunities for growth. The agent is almost like a tutor and, as people refined their responses, the coach learned and improved as well. Second, bring in an agent to help create performance summaries in seconds. The agent summarizes the feedback a person receives from colleagues throughout the year, along with metrics. That agent creates a summary of a year’s worth of input in seconds that otherwise might take half a day to synthesize. Third, offer a conversation simulator for managers, and for everyone, that lets them practice giving feedback in different situations. Through the AI agent, people experience a range of scenarios and reactions, which helps them build confidence and skill in handling difficult conversations.

I’ve seen dramatic results. These include an 89 percent increase in feedback. Ninety-five percent of users said it saved them time, and 95 percent of the comments were rated high quality, up from roughly 50 percent before we introduced AI. A manager told me the coach finally helped him say exactly what he meant in his own words, but clearer. He said the AI did not replace his judgment; it gave him the nudge and the language to be the leader he wanted to be. That is the goal: AI augmenting human intelligence and elevating what people can do.

Many other core HR areas are ripe for AI-driven transformation. Talent acquisition and mobility is one high-value area. You can let AI handle the basics such as screening support, matching, and interview scheduling, so recruiters can be true talent advisors and spend their time interacting with candidates. Guardrails are important: AI supports the process, but people make the selection decisions, with bias checks and human review throughout.

Learning and reskilling is another foundational area ripe for transformation. AI can infer skills, experiences, and role descriptions from incoming résumés and from existing people in the organization. Organizations can then build skill profiles generated by AI algorithms and confirmed and refined by people themselves. I have seen 90-plus percent accuracy in this application. AI starts it, doing the bulk of the work; people then refine it from there. That is a game changer for reskilling and for creating personalized learning paths targeted to people in roles or with skills adjacent to new ones in demand.

Many organizations have moved to new ways of working through HR Assist or “Ask HR” solutions with AI agents that respond to routine employee questions, route them to resources, reduce the time to address those requests, and free up HR business partners for more strategic support. I have seen solutions like these make a material impact, with more than a 30 percent reduction in HR costs, alongside a better experience for people.

Momentum is also building around custom GPTs and people creating their own agents to help with specific tasks. I created one for myself that acted as a CHRO flash report, pulling together and synthesizing key talent data and flagging items based on thresholds I gave it, so I always had the latest at my fingertips. I would call out AI-supported coaching as well. Some offerings apply neuroscience and ask powerful questions tailored to individual needs and role-specific needs, and they can act as tutors in a learning path.

If a CHRO is launching an AI-driven workforce transformation, where should they start with reskilling and redeploying talent, and what practical steps would you focus on in the first 12 to 24 months?

I would start with the work, not the org chart. In the first 12 to 24 months, really understand what work is being done, how it is being done today, and where the friction is. Identify and eliminate inefficient processes. Automating a broken process adds zero value.

Next, look at how the process and the tasks that make up the work can change to be more efficient, more impactful, and higher value. Tackle the highest value ROI use cases first for your business model and supporting talent needs. That gives you early proof points and momentum, and it focuses your energy where it matters most to the business.

As you map out the new work and workflow, the skills needed become clear. That enables decisions on what needs to be done by humans and what by AI and agents, where AI and agents will superpower humans and add the most value. From there, reskilling, redeployment, and hiring become much more targeted, because they are grounded in a clear picture of the reimagined work.

If AI allows companies to do far more with far fewer early-career roles, do you see a real risk of a “lost generation” of talent or a demographic gap in the workforce? What responsibility do CEOs and CHROs have to anticipate and manage that, beyond just their own P&L?

You raised the idea of a “lost generation,” and I think it is a great question. Entry roles are certainly changing, but I do not see a lost generation. Entry and early-career talent are already using AI in the flow of their schoolwork and daily activities, so I see them very naturally adapting to this changing landscape. In Gen Z new hires, they are built to learn and are growing up collaborating with agents.

Because of that, I think we need to redesign, not remove, entry-level roles. For all the reasons you mentioned, they are critical. We need people who can orchestrate systems, steward agents, and build domain depth, which takes time, and we need them to get exposure and practice applying and growing those human skills that are so important as they progress in their careers and become future leaders.

The health of the talent pipeline really depends on continuing to invest and looking at how we redesign entry-level hiring to both leverage the strengths people are already coming in with and ensure they are getting opportunities to build domain depth and human skills. You also create great opportunities for talent creation when organizations widen the aperture of access and equity through skills-based hiring programs. Many roles do not require a four-year degree if the critical skills, or proximate skills, are there or can be trained through an apprenticeship or certification program. Bringing early talent in through those channels is a very effective strategy. You can get that talent in earlier, and they can start building domain expertise and those human skills that will help them be the future leaders we are going to need.

If you were speaking to a room full of board members and CEOs, what is the one thing you would want them to know about the future of work with AI? What do they need to grasp about how workforce design, hiring, leadership pipelines, and the way they manage talent are going to change in the age of AI?

I would tell them that AI is not something you install; it is a new way of working. You earn enterprise-level value when you redesign work, rewire the operating model, and invest in people at the same pace as you are investing in technology. We have data showing that many organizations are investing in technology at about three times the rate of their investment in people. That is a miss.

I see five things you can do that really increase your odds of enterprise-level results. First, lead with value: prioritize the high-impact workflows and measure the outcomes in the P&L. Second, reinvent talent and ways of working: adopt one skills language, design roles for people and agents, and support talent pathways. Third, build an AI-enabled digital core: an integrated set of skills, data, and workflows where you can scale multi-agent systems and have appropriate guardrails in place. Fourth, operate with responsible AI: have central governance, audits, bias controls, and humans in the loop. And fifth, drive continuous innovation. This is not a project; it is how you run a company.

As we have discussed, the role of HR is really central in all of this. HR leaders will play a critical role in guiding both the organizational and cultural change that must go hand in hand with the technology.