Direction is the foundation of automotive travel. And Love’s Travel Stops, a company that has welcomed weary American road warriors for more than 60 years, knows where it’s going. It’s family-owned but has achieved massive scale (11th on the Forbes 2024 list of the largest privately owned companies in the U.S.). Its bold ambitions are matched by those of its automation team. The small unit, led by Leslie McCalla, Senior Manager of Intelligent Automation, is pushing a legacy architecture into the future. And, like travel itself, McCalla says scale doesn’t require size—it requires direction.
McCalla oversees robotic process automation across a retail network spanning 42 states and sits at the center of a multiyear modernization effort that moved Love’s from on-prem infrastructure to the cloud, expanded into citizen development, experimented with agentic automation and orchestration, and kicked off testing on IDP and LLM-driven workflows. Her perspective is grounded in execution, not unearned enthusiasm, and shaped by years inside accounting and automation.
“I’ve worked in RPA since 2017,” McCalla tells Automation Today. “Love’s already had a strong program when I joined in 2021, but we saw opportunities in scalability and we knew cloud had to be part of the strategy. We want to be right there with everybody else next year.”
Three Developers and a Ticking Clock
McCalla describes the architecture shift as the single most necessary step in expanding capability. Love’s ran more than 120 active automations at the time, built over years and dependent on on-prem infrastructure. The concern wasn’t technical feasibility—it was impact, downtime, and sheer volume.
She explains that planning was extensive, but the execution itself was faster than expected. After months of design work, all robots were migrated within weeks, including a blackout freeze during which no new development occurred.
“We were actually able to do the full migration within six weeks,” she says. “We kept the on-prem system as our fallback plan, but we decided to push forward because the future is cloud.”
Her team comprised three developers, two analysts and McCalla herself. Capacity was one of the driving reasons for change. New capabilities require bandwidth, and automation teams rarely get more headcount before they prove the need.
Citizen Development Flourishes
With orchestration modernized, McCalla’s team began offloading simpler workflow automation into the business. The goal was not democratization for its own sake. It was to free the CoE to focus on agent-assisted automation, testing new tooling, and designing for scale.
The Citizen Developer program—one of two major initiatives Love’s implemented with the assistance of partner UiPath this year—launched in January with two participants.
“We thought of it as ‘start small, start safe’,” McCalla says. “We had extra licenses, and if we did this right, we could get value without additional spend. So, we worked with the customer success team again, established governance for the program and launched into that first pilot.”
Three cohorts later, the program has 14 participants, 25 automations built, and an estimated 9,000 hours in time-savings. Use cases appeared first in indirect tax workflow filing, before expanding into fixed assets, SAP basis and card payment processes. All because Love’s had extra licenses that weren’t being put to use, and the automation team itself was running at capacity.
“Sometimes value isn’t only in dollars or hours,” she says. “Sometimes it’s a learning opportunity that teaches us something we apply elsewhere.”
Two Paths, One Use Case
With citizen development taking some of the load off McCalla’s team, they were freed up sufficiently to experiment with agentic AI. Love’s started by networking with other companies located in their home base of Oklahoma City, hosting a learning event that kicked off their agentic journey.
The team then moved on to its first agent-driven proof of concept focused on treasury wire validation—extracting account details from invoices, comparing data against internal systems, and routing final approval to a human before release of funds.
“We wanted to compare an agent-driven flow versus UiPath Maestro,” McCalla explains. “We needed to understand how to design for accuracy while keeping cost manageable.”
They built and tested both a traditional RPA with an embedded agent and one orchestrated using Maestro. They evaluated token usage, model performance, page volume, iteration cost, and error bands to determine scalability thresholds for each design philosophy.
Work continues as the team explores similar workflows using IDP. McCalla states that alternate approaches are evaluated not for replacement, but to understand which architecture scales most effectively over time.
Looking Ahead: Scale, Not Comfort
Since 2018, Love’s Intelligent Automation program has delivered sustained business value. Love’s tracks more than 146 automations, with citizen development and agentic capabilities layered on top of legacy builds. McCalla notes that automation supports growth without significant additions to staffing and, most importantly, frees people to focus on higher-value work, which improves job satisfaction.
“We want to drive value across all 600+ locations and the corporate office,” she says. “Mostly, we’re just really excited to try all these things that weren’t possible that long ago. We have a lot of energy, the right leadership and the right partnerships. We’ve been able to move fast this year.”
For next year? McCalla wants more workforce builders, more agent-driven orchestration and more model testing. These are not destinations. They are an expanding map of what automation can become when a team is driving in the right direction.
