Many organizations are experimenting with AI, but fewer are building the governance, data foundations and change management required to scale it responsibly. At ETS, that means taking a pragmatic approach, starting with real workflow challenges, not the technology, and applying AI only where it drives measurable value for teams. We spoke with Anu Miller, Head of Data and Analytics at ETS who supports our enterprise-wide AI strategy, about how ETS right-sizes solutions, measures AI impact in human terms such as time saved and capacity freed and builds trust through responsible AI practices.
When people talk about “responsible AI,” what does that look like in practice at ETS?
Miller: At ETS, we embed responsible AI practices from the start, and one way we operationalize that work is through DARES, our enterprise program which stands for Data and Augmentation with Robotics for Effective Scaling. DARES works across departments in waves to identify inefficient processes and improve them using the right level of automation or AI. Across DARES and all our AI solutions, we focus on transparent use cases, early risk assessment, bias and fairness checks and strong data privacy to bring clarity and speed to tasks. We also continuously monitor performance, cost and risk and ensure every AI solution has clear ownership and guardrails. In practice, that day-to-day work includes ongoing oversight: as models and agents are deployed, teams must manage and monitor how they behave over time, not just at launch.
That operational discipline is matched by an equally intentional focus on adoption.
To scale adoption, we partner closely with ETS’s enterprise-wide AI Champions program, which builds AI fluency through hands-on practice. This community of employees help surface local opportunities, encourage experimentation and support problem-solving across the organization. By combining strong governance with practical learning and shared ownership, ETS is working to build trust and scale AI responsibly.
How does your team decide when AI is the right solution or when a simpler approach might be better?
Miller: We start with the problem, not a specific technology, and look for the simplest approach. If we determine that a predefined process or rules-based workflow can solve the problem reliably, we take that path.
More advanced AI, such as multi-agent solutions, is used where it can meaningfully improve accuracy, efficiency or insights at scale. We think in terms of a “menu” of automation and AI options, matching the solution to the problem that balances complexity, speed and cost.
That “menu” mindset is intentional: ETS aims to right-size the solution, recognizing that not every workflow problem should be treated like an agentic AI problem.
How is AI changing the way work gets done at ETS, and what does that mean for our workforce over time?
Miller: AI helps us streamline routine work, improve decision quality and move faster on high value work that will drive business results. Over time, that means our workflows evolve and some tasks may be optimized to reduce manual effort. We are equally focused on upskilling, so our people keep pace with the technology being introduced. The goal is not just efficiency, it's enabling our teams to focus on more strategic, meaningful contributions.
How do you measure the impact of AI in human terms, like time saved, workflow improvements or capacity freed?
Miller: With DARES and our internal AI productivity efforts, we measure success in terms of time returned to teams, reductions in manual effort and improvements in decision quality and speed. Just as importantly, we look at how work changes, whether teams have more clarity, less friction, greater capacity to focus on what matters most and improved employee experience. We also consider growth in AI fluency because sustainable impact comes from teams not just using AI, but understanding how to apply it effectively
When ETS evaluates opportunities, the goal is to connect automation to real value. For example, reducing repetitive work that takes several hours each week can translate into meaningful time returned to teams over the course of a month.
What lessons from ETS’s internal AI journey would be most useful to other mission‑driven organizations?
Miller: Start with your mission and anchor every AI investment to it. This keeps priorities clear and builds trust. Just as important is investing early in governance, data readiness, upskilling and change management because real sustainable impact comes from how people adopt and use AI, not just the technology itself. For us, focusing on intentional augmentation and measurable business and human-centered outcomes has been key to scaling AI responsibly.
At ETS, governance is treated as operational work, not a one-time exercise. That includes cross-functional review of AI use cases and close partnership with compliance to help ensure AI efforts align with evolving expectations and requirements.