- For AI to be effective it can’t simply be adopted — it must be operationalized within business processes.
- While gen AI has helped drive awareness and assist with automation, agentic AI is enabling decision velocity and exponential efficiency, ushering in the next phase in business value.
- There is no perfect time to start using agentic AI. Make it a priority to incorporate it into your business processes now so agents can learn and grow alongside your organization.
As AI solutions flood the HR market, it’s easy to get lost in the fanfare. The organizations that will get ahead are the ones who look beyond the buzzy potential to determine which technology is right for their organization and then actually put the technology to work.
That was the main message that R “Ray” Wang, Principal Analyst, Founder, and Chairman of Constellation Research, shared in his keynote at Cultivate ’25.
“It’s not about technology hype,” Wang said. “It’s not about all the possibilities that might happen. It’s about all the decisions that you’re going to want to put in one place, and organizations that actually do this that are going to have that winning business model going forward.”
To emphasize that point, Wang showed an old photo of a refrigeration system as an example of a “hot new technology” that was forecasted to transform businesses operations in a past era. He asked the audience who they thought the winner was in moving the technology forward.
The answer: The Coca-Cola Company, because of the way the enterprise established cold supply chains and built the technology into their day-to-day work.
Similarly, in this new era of agentic AI, Wang says, enterprises that successfully integrate the technology into their overall strategy and find ways to complement the work of humans will come out ahead.
Here are eight insights Wang shared about putting AI to work.
Check out some of the highlights from R “Ray” Wang’s keynote on buying agentic AI at Cultivate ’25.
1. In this age of margin compression, AI will enable every employee
Nearly every organization is facing pressure to drive down costs while improving margins, and it’s getting ever more challenging to do so among widespread layoffs.
But there’s a side of the story that too often goes untold, Wang says: “Some people are getting a lot more done with the talent that they have.”
That’s what happens when every employee is enabled by AI: they’re empowered to work better, whether they’re in marketing, IT, HR, or any other department.
Wang says organizations can benefit from identifying highly repetitive efforts, automating them and then finding opportunities for talent to figure out “the next big thing.” AI can also enable organizations to serve consumers better and differentiate customer experiences.
2. When used correctly, AI takes over tasks and elevates human experiences
Recently, Wang had the opportunity to interview Chris Nicholas, President and CEO of Sam’s Club, about AI implementation, which allowed the business to eliminate 100 million tasks.
While you might expect that that would lead to layoffs, Wang reported that Sam’s Club actually hired 3-4% more workers on the frontline so they could talk to customers, guide them through stores, and answer questions.
“The story isn’t necessarily what everyone imagines it to be,” Wang said. “You’re getting efficiencies, you’re transforming employee experiences, and you’re putting back mission and purpose.”
Wang explains how the 10x rule applies to where AI is heading right now.
3. Agentic AI is the next evolution
Gen AI has generated a lot of buzz since it became widely accessible to the masses. It’s helped individuals complete their homework, write speeches, revise résumés, and more.
“It was pretty cool,” Wang said. “But gen AI doesn’t necessarily solve a problem, in terms of an outcome or action.”
That’s where agentic AI comes in.
“Because it’s enterprise-ready, it means you’re figuring out what step of the process — from hire to retire — that you’re automating, and you’re going to put an agent in place to make decisions on your behalf, while making sure a human is in the loop to make sure something doesn’t go wrong,” Wang said.
Agentic AI will lead to decision velocity, he says, adding that agents will pave the way for “exponential efficiency,” completing work in a manner that’s 10 times better, 10 times faster, and 10 times cheaper than humans.
4. Data is AI’s fuel (and you probably don’t have enough of it)
To reach that decision velocity and exponential efficiency, organizations need data from several sources. Wang says there’s an unofficial understanding that 85% accuracy in data is good enough.
He went on to say that in some cases that may be true (customer experience and employee experience), while in others it’s simply not sufficient (supply chain, finance, and health care).
To enable agents to be more effective, organizations must augment existing data by 10% with information from inside the business (including transactions, HR data, CRM data, and payable information) while 5% of data should come from partners, signals, or outside sources.
“We don’t have enough data to get to a level of precision we trust; 85% is not going to work,” Wang said. “The more active data signals you put together, the smarter your agents will be. With smarter agents we can do a lot more.”
Wang talks about the importance of humans and AI working together in the autonomous decade.
5. People must be in the loop
There’s no question that businesses can automate a whole host of processes, and agents will continue to learn and fine-tune abilities with time.
But we can’t underestimate the value of keeping people involved in the work, whether it’s a physician diagnosing and explaining a health condition to a patient, or a pilot responding to a flight emergency mid-air.
It’s essential to blend our abilities with machine capabilities so that each complements the other. Businesses that do that, Wang says, will reap the benefits.
6. Future managers may oversee agents alongside people
“This may be the last generation of managers that only manage humans,” Wang said, adding that managers of the future may also coordinate fleets of agents alongside people.
Those agents may start out slow, he warned, but with training they will get smarter and better over time.
“We started out with this notion that we’re going to augment and bring that to life. And then we took that augmentation and realized, ‘Oh, wait, we can accelerate some of these business processes,’” Wang said. “Now we’re looking at automation agents. The next step is the advisers that actually start making those decisions on their own.”
When that happens, it will fuel a major shift in organizational design.
7. Business leaders must consider risk, compliance, and ethics early and often
Before deploying agentic AI, organizations need to think through all possible scenarios and have contingency plans to protect the business.
Wang shared the example of a bird striking an automated plane over the Hudson River. If that plane goes down, thousands of lives could be impacted. What will the machine do? And, alternatively, if the plane crashes and causes harm, who is at fault?
“Make sure it’s bulletproof, especially if it touches a human-critical system,” he said.
8. The right time is now — even if it’s not a perfect start
There’s no ideal time to implement AI solutions, and there never will be. The key is to move beyond the buzz, find the right solution for your business and get started.
Along the way, keep training your agents and guide your organizations from augmentation to automation to full-scale decision-making support to enhance your HR journey.
“To attain business value and outcomes with AI,” Wang said, “the first thing you have to do is get started.”
Watch Wang’s Cultivate ’25 session, “The buyer’s guide to agentic AI: How to determine which agents are right for you,” available now on demand.