The Identify phase of Tarpon Health's Automation Lifecycle exists for a simple reason: not every process deserves automation. Some workflows look promising on paper but fail to deliver meaningful returns. Others seem too complex but actually offer significant value once you dig in.
Skipping or rushing this phase leads to common problems. Teams build automations that nobody uses. Projects stall because stakeholders weren't aligned from the start. Worse, organizations automate processes that shouldn't exist in the first place, cementing inefficiency rather than eliminating it.
A structured identification process helps you avoid these pitfalls. You'll spend resources on projects that actually move the needle, and you'll build organizational buy-in along the way. The time invested here pays dividends throughout the entire automation lifecycle.
01 Educate Your Stakeholders
Before you can collect automation ideas, stakeholders need to understand what makes a good candidate. This includes business process owners, subject matter experts, and management at various levels.
Good automation use cases typically share three characteristics:
Digital interaction. Staff work within software systems, clicking through screens, copying data between applications, or navigating portals. Manual paperwork or phone-heavy processes require additional steps before automation becomes feasible.
High volume and repeatability. The process happens frequently enough to justify development effort. A task performed twice a month rarely warrants automation, while something happening hundreds of times daily almost certainly does.
Limited outcomes. The process follows predictable paths with a finite number of results. If every transaction requires unique judgment calls, automation becomes difficult. But if 80% of cases follow the same pattern, you have a strong candidate.
Address concerns early. Staff worry about job security when automation comes up—that's natural. Be transparent about how automation typically shifts work rather than eliminates it. Most organizations find that bots handle the repetitive tasks while humans focus on exceptions, complex cases, and patient interactions that require judgment.
02 Collect Ideas From Your Team
The people doing the work know where the pain points are. They understand which tasks consume disproportionate time, which processes break down regularly, and which workarounds have become unofficial standard practice.
Several approaches work well for gathering ideas:
Roadshows and demos. Visit departments with presentations showing what automation looks like in practice. Concrete examples spark recognition: "We do something just like that."
One-on-one meetings with directors. Department leaders often have a broader view of workflow bottlenecks and can identify patterns across their teams.
Discovery workshops. Monthly sessions where staff can propose and discuss potential automation candidates. These work best when participants see that previous suggestions led to actual projects.
Structured intake forms. Create a simple way for anyone to submit ideas at any time. A consistent format makes evaluation easier and signals that you're serious about considering all suggestions.
The goal isn't just collecting a list of processes. You're building a culture where people actively look for automation opportunities and feel invested in the program's success.
03 Pull and Analyze Key Data
Data reveals what conversations might miss. Objective metrics help you identify high-impact opportunities and build compelling business cases.
Focus on indicators that highlight automation potential:
Process volume. How many times does this task occur daily, weekly, or monthly? Higher volume typically means higher automation value.
Denial rates by category. Which denial types consume the most rework effort? Patterns in denial data often point to upstream process failures that automation can address.
Write-off trends. Where are you losing revenue? Timely filing limits, missed coverage, and uncaptured charges all represent automation opportunities.
Backlog depth. Which queues consistently overflow? Persistent backlogs indicate processes that can't keep pace with volume—prime automation territory.
Time-to-complete metrics. How long does each step take? Processes with highly variable completion times often have steps ripe for standardization and automation.
This data serves multiple purposes. It helps you identify candidates, yes, but it also establishes baselines for measuring automation success later.
04 Shadow the Process
Reading documentation and talking to staff provides valuable context, but nothing replaces watching the work happen. Process shadowing reveals the reality beneath the official workflow.
Sit with the people who perform the task. Watch their screens. Note where they pause, where they switch between systems, and where they apply workarounds that aren't documented anywhere. Ask questions as they work: "Why did you do that step? What happens if this field is blank? How often does this exception come up?"
Document what you observe:
- The actual sequence of steps (which often differs from documented procedures)
- Systems and screens touched
- Decision points and their criteria
- Common exceptions and how staff handle them
- Time spent on each segment
This shadowing data feeds directly into feasibility assessment. Can a bot navigate these systems? Are the decision rules codifiable? Does the process depend on institutional knowledge that would be difficult to replicate?
Equally important, shadowing validates volume estimates. A process that theoretically handles 500 transactions daily might actually process far fewer, or far more, depending on how work flows through the system.
05 Prioritize Strategically
With a list of potential automations, you need to decide where to start. Prioritization ensures you pursue opportunities that balance impact with feasibility.
Two factors matter most:
Return on investment. Calculate expected value by combining volume, time savings per transaction, and labor costs. Factor in error reduction and downstream benefits like fewer denials or faster collections. Higher ROI candidates deserve priority.
Complexity. Assess technical difficulty, number of systems involved, exception handling requirements, and stakeholder alignment. Lower complexity means faster implementation and higher success probability.
Plot your candidates on a simple matrix. The sweet spot combines high ROI with low complexity—these are your quick wins that build momentum and demonstrate value. High ROI but high complexity projects become your strategic initiatives, worth pursuing but requiring more careful planning.
Avoid the temptation to start with complex, high-profile projects just because they're visible. Early wins matter for program credibility. A dozen successful small automations do more for organizational buy-in than one ambitious project that drags on for months.
Building Your Automation Pipeline
Identification isn't a one-time event. The most successful automation programs maintain a continuous pipeline, always cultivating the next round of opportunities while current projects progress through development.
Schedule regular discovery sessions. Review operational data quarterly for emerging patterns. Keep intake channels open and responsive. When staff see their suggestions turn into working automations, they'll bring more ideas.
The pipeline approach also helps with resource planning. You'll always know what's coming next, which makes staffing decisions and vendor conversations more productive.
Frequently Asked Questions
Q: What is RCM automation?
RCM automation uses software bots and AI to handle repetitive revenue cycle tasks like eligibility verification, claim status checks, and payment posting without manual intervention.
Q: How do I know if a process is a good automation candidate?
Look for high volume, digital interaction, and predictable outcomes.
Q: What's the biggest mistake organizations make when starting automation?
Skipping the identification phase and jumping straight into development without validating that the selected process will deliver meaningful value.
Q: What data should I analyze to find automation opportunities?
Process volume, denial rates, write-offs, backlog depth, and time-to-complete metrics reveal where automation delivers the greatest impact.
