Prioritize Your Fires

Elise Miller is the VP of Customer Experience at Daloopa, the platform powering public equity professionals via trusted financial data and AI automation across the investment research cycle. Before Daloopa, she spent seven years at WalkMe, rising from CSM to Director of Customer Success across mid-market and enterprise segments. She studied psychology at Brown University. And before any of that, she was a professional ballet dancer.

That last detail isn’t a fun fact. It’s the key to how Miller thinks about AI in customer success. A career built on discipline, reading subtle cues, and understanding what drives human behavior has given her an instinct for the signals that don’t show up in dashboards. In fintech, where customers are too time-crunched to tell you what’s wrong, those signals are everything.

From the Stage to the Data

Miller’s path into tech started with a pivot most people wouldn’t expect. After training her entire life for professional ballet, she joined a company, and after a year, realized the job version of her dream wasn’t the dream itself. She enrolled at Brown, chose psychology, and found herself drawn to courses on human behavior, decision-making, and digital behavior.

The connection to customer success clicked almost immediately. CS, she says, is “a lot of relationship management and really understanding people, which is just as important as being data-driven and a product expert.” That blend of empathy and analytical rigor carried her through seven years at WalkMe, where she managed everything from mid-market accounts to strategic enterprises and grew into a leadership role overseeing services and success teams across the East region.

When Daloopa came calling, what attracted her wasn’t just the fintech space or the AI angle. It was the chance to build a customer success function from scratch at a company where the customer base was only beginning to embrace the AI curve. That same instinct for reading people now runs on a very different kind of stage.

What’s Actually Working

Daloopa serves roughly 185 accounts and thousands of individual users, most of whom are analysts and investment professionals who barely have time for a quarterly sync. Getting customer sentiment the traditional way is a nonstarter. “We actually have to fight for our customers’ attention,” Miller says. That constraint forced her team to get creative with data.

The most impactful experiment has been using AI tools to diagnose why users go inactive, not just flag that they have. Miller’s team ran analysis on a recurring basis, analyzing what behaviors users showed leading up to them becoming dormant. Did they hit a bug? Try a feature that wasn’t mature enough? The goal isn’t a count of inactive users to hand to product. It’s a hypothesis for each one.

“We’re becoming a much better engine for our product team, because we’re going so much deeper. It would’ve taken us months to find the leading indicators of someone becoming inactive — now we can query that quickly and easily. That’s been one of the biggest things: using AI to get to the root cause of inactivity, rather than just accepting that these people are inactive.” — Elise Miller, VP of Customer Experience at Daloopa

The results have been concrete: inactive user rates have dropped significantly, and the team is reactivating more users than at any point in the company’s history. Miller ties it directly to retention and expansion, noting that inactivity and churn are tightly correlated in their business.

She’s also built AI into how she coaches her team. Using Outreach IO, her CSMs follow structured re-engagement sequences for dormant users. In finance, phone calls are how you actually reach people, so the team’s playbook is call-heavy by design. Miller ran an analysis that surfaces each CSM’s outreach mix and scores quality against a rubric: Are they setting agendas on calls? Using the approved email templates? Some CSMs were over-indexing on email and not making enough calls, while others weren’t reaching out at all. The analysis turned vague coaching conversations into specific, data-backed ones, letting Miller and her managers focus their time on the adjustments that actually moved the needle.

When the Model Hallucinates Your Champions

Miller’s relationship with AI wasn’t always this refined. One of her earliest experiments was a cautionary tale she still tells her team. She wanted to mine Gong recordings for customers who might provide positive testimonials about a new product. She got back a list she was excited about.

Then she looked closer. One “champion” had simply expressed interest in trying the product, not endorsing it. Another was an internal salesperson talking about a customer in the third person on a team call. A third was pulled from context that had nothing to do with genuine advocacy.

“For AI to work properly, you have to actually really understand what your end goal is and be very, very specific. When you’re vague or subjective with the prompting, it can seriously affect your workflow in a negative way.” — Elise Miller, VP of Customer Experience at Daloopa

The team didn’t end up reaching out to any of those false positives, but the lesson reshaped their entire approach. Now the CS team maintains a repository of dashboards and analytics queries the team has built, tested, and confirmed produce reliable outputs. If someone ventures outside the verified set, the expectation is simple: check a few data points before you run with it. Inside the set, trust the output and move fast.

Beyond the At-Risk User

Miller’s next move is to stop treating AI-driven outreach as purely reactive. Right now, most of her team’s workflows target users who are already in trouble: dormant, flagged by declining usage, or coming off a bad product experience. The bet for the next two to three quarters is to expand that lens to every user, pulling in web-based news and persona-level context to send proactive, personalized outreach that isn’t triggered by a red flag.

She’s also pushing to build a true health score, one that stitches together usage patterns and data into something predictive rather than descriptive. The pieces exist. They just haven’t been wired together yet.

“Prioritize your fires. AI is a great tool, but when you try to fix everything at once, you kind of end up fixing nothing.” — Elise Miller, VP of Customer Experience at Daloopa

The teams that win won’t be the ones running the most experiments. They’ll be the ones who finished what they started.

The Sonora Project captures how CS leaders are navigating AI — the real experiments, practical wins, and hard-won lessons shaping the future of customer success.