Many practice owners feel pushed to make technology decisions before they've had a chance to look at the work itself. In this episode of The Thriving Practice Podcast, Tracy Cherpeski talks with Denisse Rodriguez, founder of InService AI in Toronto, about how to improve operations in a healthcare practice before choosing an AI or automation tool.
If you're a practice owner feeling the weight of administrative work, or you've been tempted by a slick vendor demo, this conversation offers a calmer way forward. Denisse explains why a practice that grew organically often ends up with disjointed processes, and why mapping your workflows first is the most reliable path to a more profitable, less exhausting practice. It's a practical look at how to run a successful practice and protect your team from burnout without adding another shiny tool to the pile.
Tracy and Denisse share a core belief: slow down before you speed up. You'll hear how to tell whether you're facing a people problem, a process problem, or a tool problem, and how to define success in terms of both measurable results and the way your team actually feels.
Key Takeaways
Diagnose before you automate. If your workflow has holes, speeding it up with AI just helps you get nowhere faster. Fix the leaks first.
Define outcomes in two ways. Set tangible goals like a better intake or faster patient communication, and intangible ones like less frustration and a better experience for patients and staff.
Watch for "transferable time." A tool that makes you spend the same hours validating its output hasn't saved you anything.
Start small and measure. Run small, manageable tests, measure before and after, and be willing to adjust quickly if it isn't working.
Ask vendors hard questions. Security comes first, then whether a demo shows what's available now or what's coming later.
Q&A
Why should practice owners fix workflows before adopting AI?
Denisse uses a bucket analogy: pouring more water into a bucket with holes gets you nowhere. Practices tend to grow organically, so processes become inconsistent over time. Taking the time to review how work actually happens lets you decide what to keep, what to fix, and what to automate.
How do I know if I need AI, more staff, or better processes?
Tracy suggests asking whether you're solving a people problem, a systems problem, or a tool problem. Denisse notes that the symptom, like hiring challenges, often masks the real issue, which is how capacity is being lost in current processes. Writing down how you spend your day is a good place to start.
What should I ask a vendor before buying an AI tool?
Denisse says to start with security, which is a deal breaker in healthcare if it isn't there. Then ask whether what you saw in the demo is available today or is a future release, and whether the tool will serve your specific objective rather than just looking impressive.
Which tasks are best to automate first?
Denisse recommends looking at tasks that are repeatable, high in value, and low in judgment. Work that has to get done but doesn't add much value in itself is a strong candidate, which frees your people for the work that needs a human.
Episode Highlights
How InService AI helps organizations improve operations using AI and automation
Why many practice owners make technology decisions before examining their workflows
The bucket analogy: why automating a leaky process doesn't help
How a symptom like hiring struggles can mask an operations problem
Why there are around 40,000 AI tools, and why starting with an assessment matters
"Transferable time" and how a tool can shift your hours instead of saving them
A credentialing example: how resistance and poor data made an automation output useless
Why measuring before and after builds momentum for your next use case
Tangible vs. intangible outcomes and why both matter
Using your clinician hat and CEO hat together to diagnose business problems
Small iterations: test small, fail quickly, recover quickly
Smart questions to ask vendors, from security to demo reality checks
Denisse's background in Mexican folklore dance and what choreography has in common with operations
Memorable Quotes
"You could keep throwing water into the bucket, but if it had holes... you're going nowhere." - Denisse Rodriguez
"If you are not clear on where you're going, it will get you there faster, which is where you don't want to go." - Denisse Rodriguez
"You just transferred your time from actually reading to validating the output." - Denisse Rodriguez
"Small iteration, fail quickly, recover quickly." - Denisse Rodriguez
"Call me crazy, but I find beauty in the operations world." - Denisse Rodriguez
Denisse's message is simple: get honest about your workflows, get clear on the outcome you want, and let the technology follow. When you slow down enough to see the leaks, the right solution becomes much easier to spot. If this conversation sparked ideas, share it with a colleague who is considering AI for their practice. Connect with Denisse on LinkedIn at InServiceAI, and explore thrivingpracticecommunity.com for more resources to help you build a practice that thrives.
Guest Bio:
Denisse Rodriguez spent 30 years as the person global organizations called when data, technology, and people needed to work together at scale, leading programs across more than 70 countries, many of them at Deloitte. Along the way she noticed that the organizations doing the most meaningful work often had the fewest systems to support it. That led her to found In Service AI in Toronto, where she helps teams cut through operational overwhelm with practical, human-centered AI. She also studied Mexican folklore dance and believes a well-coordinated operation is a thing of beauty.
Find Denisse:
Connect With Us: