How Do You Prove AI ROI When The Return Is Only 1%?

Most companies are still figuring out what an AI agent should actually do. JLL has moved well past that question. With agents operating across facilities management, brand communications, and corporate functions, the company is learning that the hardest problems aren’t necessarily technical—they’re about proving value, building trust, and making sure AI solves an actual business problem instead of becoming another expensive corporate science project.

Recorded at the AI4 conference in Las Vegas—with some of the inevitable conference-floor background noise—this conversation between host David Rice and Carlin Power, Global VP of AI and Platform Technology Enablement at JLL, explores what it actually takes to deploy AI at scale. They discuss why ROI needs to live inside the P&L, how decentralized AI development can accelerate useful innovation, and why eliminating entry-level roles in the name of efficiency may be exactly the wrong lesson to take from AI.

What You’ll Learn

  • Why successful AI agents need to connect directly to measurable business outcomes
  • How a strong enterprise data foundation makes decentralized AI development possible
  • Why AI fluency requires more than simply giving employees access to tools
  • How augmentation can raise expectations for early-career talent rather than eliminate those roles
  • Why leadership behavior and incentives matter when trying to drive genuine AI adoption
  • How organizations with no agents in production can start small without getting stuck in permanent pilot mode

Key Takeaways

  • Put AI inside the business problem, not beside it. Asking for the ROI of “AI” in isolation is often the wrong conversation. JLL pushes the question back into the P&L: if a business unit needs to increase throughput, improve client outcomes, grow revenue, or expand share of wallet, what role can AI play in achieving that target? Otherwise, you’ve got the familiar hammer wandering the halls looking for a nail.
  • Treat agents like investments with performance expectations. JLL evaluates whether agents are delivering the benefits they were built to produce and uses a recurring stop, start, or continue conversation to determine whether further investment makes sense. An agent shouldn’t get a free pass simply because everyone currently feels obligated to put “AI” somewhere in the strategy deck.
  • Don’t confuse automation with workforce strategy. Power argues that eliminating early-career talent is a sign of an immature AI strategy. JLL’s interns are using AI to tackle sophisticated projects at a level she says can resemble employees several years into their careers. The opportunity isn’t necessarily fewer junior employees; it’s radically higher expectations for what junior employees can contribute.
  • Fluency has to be built, not announced. JLL invested in enablement professionals, training, champion programs, and role-specific education across a company operating in 80 countries. Employees learned both what the technology could do and where they still needed to “trust but verify.” Access to a tool is not the same thing as knowing how to use it responsibly.
  • Start narrow enough to learn something useful. For leaders beginning with zero agents, Power recommends assembling technically literate domain experts and identifying a small number of use cases where experimentation itself can produce positive ROI. Build evidence, scale what works, and reinvest the returns rather than trying to transform the entire organization on day one.
  • Leaders need to use the tools themselves. JLL paired bottom-up innovation with top-down role modeling, including KPIs around AI fluency, adoption, and usage. Employees are unlikely to experiment confidently if leadership treats AI like something everyone else should be doing.

Chapters

  • 00:00 — What the Other 97% Will Learn
  • 02:08 — Moving AI Agents Into Production
  • 05:40 — The AI ROI Problem
  • 07:48 — Putting AI Inside the P&L
  • 09:26 — Building Agents That Last
  • 10:43 — Scaling Decentralized AI
  • 13:37 — AI and Entry-Level Work
  • 17:09 — Rethinking Early-Career Talent
  • 19:00 — Building AI Fluency at Scale
  • 21:12 — Your First Three Months With AI Agents
  • 24:41 — Closing Thoughts

Meet Our Guest

Carlin Power is the Global Vice President of AI and Platform Technology Enablement at JLL, where she helps drive enterprise AI adoption and technology transformation across the global commercial real estate organization. With a background spanning commercial real estate, product management, analytics, technology enablement, and employee development, she focuses on turning emerging AI capabilities into practical tools and workflows that improve how people work. Carlin is a passionate advocate for people-centered AI transformation, emphasizing education, responsible adoption, and measurable business outcomes while helping employees build the skills and confidence to thrive in an AI-enabled workplace.

You may also like