Green Touch Services Client Story
Within the past few weeks, the project manager at Green Touch Services (GTS) was running a contract renewal through her new AI-assisted workflow when it flagged something that almost slipped by. Seven services on that account were running below their target profit margin, with a few of them quietly losing money for 3 consecutive years. The new workflow also caught a service efficiency anomaly and would have significantly overcharged the customer. Two services she had priced in her first, manual test draft came in below the existing contract rate, and the workflow flagged those before the proposal ever went out the door.
The funny thing is that none of that was the initial goal when they started exploring Amazon Quick. GTS initially set out to get better visibility into their data through dashboarding but ended up with a completely tailored AI workspace that is changing the way they collaborate and work.
Here’s how they got there and what that effort looked like, including the less-polished middle that doesn’t usually get shared.
Life before Quick: plenty of data, no insight
GTS is a service-based landscaping company that built a custom job-tracking application shaped specifically around how they work. They chose Claris FileMaker for flexibility and rapid development and tailored the application to their specific process, tracking costs, logging work orders, and generating efficiency metrics.
The problem was never a lack of data – it was a lack of insight into their data that they could act on.
“We knew we were making money. You can see that in your end-of-year P&L,” she told me. “But you’re not making as much as you could be, because you’re going into all these contract renewals and service agreements so blindly.”
She and her team had created a series of custom FileMaker reports. One was an efficiency tracker she knew inside and out. She could extract many of the answers that others wanted. But it was also, by her own playful admission, “hideous”. It was packed with data but nearly impossible for anyone else to be able to decipher it on their own.
The information existed in their FileMaker data and in her head, but others frequently relied on her institutional knowledge to understand it. Without accessible, visual, queryable insights, pricing decisions were made on instinct and intuition.
Jumping in with hope and a blank screen
When GTS began looking at BI tools, Tableau was on the shortlist. However, Quick soon rose as a top choice based on the AI features and capabilities that they were interested in. Turning a promising demo into deep platform expertise, though, was its own journey.
She’s the first to laugh about how green she was going in. “You couldn’t have had a greener plant,” she said. “I didn’t even know what I wanted it to do. All I know is that I wanted it to do something.”
That experience is far more common than most users will admit and it’s worth highlighting. Evaluating a new platform often uncovers a gap between demonstration use cases and a company’s reality: it’s not always apparent how you may want to engage with a new workspace and you might not have enough context to connect the dots.
Working with a consultant, a real-life use case can be discovered together, but that takes trust in the process from both sides. When the client can name one or two concrete objectives before we start, progress is much more focused and linear. As she reflected back, “The client has to have an actual objective or two, other than ‘we want to be able to massage our data better.’ That’s pretty vague.”
Three breakthroughs that changed everything
The first turning point was the day she fully built a working dashboard on her own. After the effort of cleaning up their data and hooking up a data pipeline to Quick, she sat back in revelation. “Holy smokes,” she told me. “It showed we had good data, finally.” It was proof that the upfront work was worthwhile and that Quick could deliver the insights that they wanted. For someone who didn’t have an initial use-case, that first working dashboard was the revelation that they could do more with their data and things started to click.
The second breakthrough came a short time later. She had not yet spent much time using their initial custom AI agent, “Sprout”, until one day she asked a real business question and got a real, data-backed answer in return. “That was pretty incredible,” she said. “Once you involve that, you feel like you’ve unleashed some power.”
The third came when she stopped reacting to Sprout’s insights and instead began collaborating with it, training it to assist with everyday tasks. She started with her monthly property summary (a manual report that used to take about an hour per client) and turned it into a fully automated Flow. “I really don’t have to do anything on that now,” she told me. Because of the time involved, team members historically shied away from sending routine updates, even amid increasing requests for transparency. Since some team members each manage 10 or more clients, the Flow can automate 10+ hours of work a month per person, a significant time savings.
Next, she turned towards her daily time clock audit, which took her 20 – 30 minutes each morning. Passing this off to their AI agent has already cut her time down to about 10 minutes and is next in line to become a Flow to completely automate the effort.
Furthermore, within the past few weeks, she wrote a 10-step SOP for their contract renewal analysis, a structured back-and-forth conversation between herself and Sprout that ensures every renewal is priced accurately, no problem services slip through, and leadership gets a clean summary before anything goes out to their clients. She ran a few test renewals through it, refined it each time, and then had Sprout email her team with the results, with a plan to roll it out to other reps. During this collaborative process, Sprout caught the underperforming services and data anomalies before the contract renewal was submitted to their client.
Turning revelation into operational change
This is the heart of deploying an AI workspace. A dashboard can show you what you have already asked to see but an AI agent can collaborate with you to explore what to ask. For a business owner trying to envision what to do without a precise use case, this can revolutionize how they work.
This was the shift I’d been hoping to see. Our client went from “what can this show me?” to “how do I build a repeatable process around what we do?” The journey can be the same for you, whether you have a renewal SOP, a monthly report, or a daily audit: we identify a process that depends on data, build a structured workflow around it, and hand it off to Quick so it can scale beyond one person. Then, when you’re ready, you take the leap beyond the dashboard, running your data-driven operations in partnership with an AI agent helps you investigate what’s actually happening in the present moment and take action on what you see. This means that you now have more time (and mental energy) for the work where your time and expertise matter most.
What nobody tells you about data readiness
Here’s the part that usually gets left out of the success story but that’s a fundamental step in the process. You’ve got to clean up your data.
Many organizations sit on years of flexible, organically grown data that doesn’t have the structure and cleanliness required for AI to process in a flexible and trustworthy way. This is especially true with custom applications built on permissive platforms such as FileMaker.
Flexibility is one reason that companies like GTS build on the Claris platform: it’s easy for knowledge workers to build homegrown applications customized to their business needs. While FileMaker supports strong data integrity controls, it doesn’t insist on them. If data guardrails aren’t designed properly from the start, text gets stored in numeric fields, users enter wonky dates, and exceptions to standardized values abound.
This principle extends well beyond Quick. Hand bad data to any AI tool, whether it’s an AI assistant, agent, or other application built on your data, and you’ll get bad results. The difference is that Quick dashboards tend to fail loudly as it flags what it can’t parse and makes the problem obvious. Gen-AI tools, on the other hand, fail more quietly. They’ll still give you an answer as they’re designed to be helpful, even when that answer is built on a shaky foundation. That makes dirty data even more dangerous because you might not realize anything is wrong.
The GTS work surfaced data cleanliness issues over their million-plus rows of data. Each one arrived as its own surprise, and sometimes it felt cyclical. We would clear one hurdle and then another appeared on the horizon in the next table. The frustration was real in the moment, but she tells it now as the rollercoaster it was, not from a place of regret. When I described the cleanup as data renovation rather than a coat of paint, she agreed: “Right, exactly. We weren’t just painting the walls.”
During the course of this project, I built tools that make data cleanup faster and repeatable. They assessed all her records and flagged problematic entries for cleanup in FileMaker, empowering her to pursue the cleanup on her own time, saving budget along the way.
The lesson I took for future engagements is that a data readiness assessment up front is essential, even though it’s not the most fun part of the process. It won’t catch everything, because we don’t always know which tables Quick will need as the project evolves, but it sets honest expectations and gives everyone a shared picture of what “readiness” really means. Most importantly, it shortens the path to engaging with every AI feature that comes next.
The payoff: a foundation that keeps paying
GTS’ new chat agents are insightful, their dashboards are grounding, their Flows are saving them time, and their new renewal numbers speak for themselves. GTS came out of this effort better understanding their own data, taking responsibility for its quality, and building automations and AI capabilities that catapult themselves into a new way of working. They have moved from reactive fixing to preventive stewardship, with benefits that compound over time.
If your data lives in FileMaker – or any system that’s grown and evolved alongside your business – and you want to explore what AI tools can do for you, we’d love to partner with you and share our experience. We know what the messy middle looks like and we know what’s on the other side: the ability to move beyond static reports to dig into your data, collaborate with it in real time, and do something meaningful with what you discover.
