Article
Why Ecosystem Design Controls Are Really About Moving Faster With the Right Rigor
How MedTech Teams Must Evolve for the Agentic AI Era
Agentic AI is changing how MedTech product teams build software.
We are moving beyond AI as a mere “copilot” and into an era where multi-agent systems can plan, reason, and execute complex product-development tasks. These systems can help generate requirements, draft user flows, write and test software, map traceability, support ISO 14971 risk analysis, and create regulatory documentation based on system architecture.
For MedTech organizations, the promise is significant: faster development cycles, stronger cross-functional leverage, and greater market velocity.
But there is a catch.
Many teams are not seeing the ROI they expected. They buy enterprise licenses. They provision multi-agent frameworks. Then, months later, they realize AI is being used as a glorified spellchecker or autocomplete tool.
The problem is not always the technology.
The problem is often what Larkin Lowrey, Chief Technology Officer at Orthogonal, calls IC Brain.
Agentic AI is changing how MedTech product teams build software.
We are moving beyond AI as a mere “copilot” and into an era where multi-agent systems can plan, reason, and execute complex product-development tasks. These systems can help generate requirements, draft user flows, write and test software, map traceability, support ISO 14971 risk analysis, and create regulatory documentation based on system architecture.
For MedTech organizations, the promise is significant: faster development cycles, stronger cross-functional leverage, and greater market velocity.
But there is a catch.
Many teams are not seeing the ROI they expected. They buy enterprise licenses. They provision multi-agent frameworks. Then, months later, they realize AI is being used as a glorified spellchecker or autocomplete tool.
The problem is not always the technology.
The problem is often what Larkin Lowrey, Chief Technology Officer at Orthogonal, calls IC Brain.
What Is IC Brain?
“IC” stands for Individual Contributor. IC Brain is the belief that professional value comes primarily from manually generating artifacts.
For engineers, that artifact is code. For Product Owners, it is a requirements document or Jira ticket. For SDETs, it is a test script. For Quality System Engineers, it is an FMEA or traceability matrix. For leaders, it may be the metrics dashboard, the implementation plan, or the rescue work they personally perform when a team gets stuck.
In traditional product development, organizations often rewarded the people who could produce these artifacts fastest and most meticulously. But Agentic AI changes the value equation.
When AI can generate the first draft of code, requirements, tests, risk documentation, and traceability, the highest-value work shifts from manual production to orchestration.
The expert’s role is no longer to type every artifact by hand. It is to define the constraints, provide the clinical and regulatory context, audit the machine-generated output, and remain accountable for whether the system is safe, compliant, and commercially useful.
The danger comes when experienced professionals push the AI aside because its output does not match exactly how they would have done the work themselves.
They think:
“It’s just faster if I do it myself.”
That instinct is understandable. It is also the bottleneck.
The Five Role Shifts Behind IC Brain
IC Brain does not show up in just one function. It appears across the entire MedTech product-development lifecycle: engineering, product, testing, quality, and leadership.
That is why Agentic AI adoption cannot be treated as a simple tooling upgrade. It requires an operating model transformation. Teams must shift from manually generating artifacts to orchestrating digital labor through better constraints, richer context, stronger review practices, and clearer accountability.
The full Orthogonal white paper, How MedTech Teams Must Evolve for the Agentic AI Era, explores these five role shifts in detail.
How IC Brain Shows Up Across the Product Team
1. From Coder to Agentic AI Orchestrator
For software engineers, the shift is from writing every line of code to defining the contracts that govern code generation.
The Orchestrator Engineer does not begin by asking AI to “write a script.” Instead, they define schemas, interfaces, memory constraints, performance expectations, security requirements, and architectural boundaries.
The key shift:
Write contracts, not logic.
This does not make engineering expertise less important. It makes architectural judgment, security acumen, and human accountability more important. The engineer becomes the reviewer and governor of the system, not merely the fastest person at the keyboard.
2. From Spec Writer to Agentic AI Orchestrator
For Product Owners and Product Managers, the shift is from dictating screens to defining clinical intent.
A traditional product workflow may start with a specific UI request: add this button, place it here, open this modal, export this file. In an agentic workflow, that level of prescription can limit the AI’s ability to reason about the best solution.
The Agentic Product Owner defines the user goal, clinical workflow, business rules, and regulatory constraints.
The key shift:
Define the clinical need, not the screen.
This moves product work away from backlog administration and toward strategic orchestration of user value.
3. From Test Scripter to Agentic AI Orchestrator
For SDETs, the shift is from writing automation scripts to governing behavioral contracts.
If AI can generate tests, the SDET’s highest-value work is not manually writing every selector or maintaining every script. It defines what the system must prove, where it might fail, and how verification should map to clinical and regulatory risk.
The key shift:
Manage the contract, not the selectors.
In MedTech, this matters because verification is not just a technical step. It is part of proving that software is safe and effective.
The Agentic SDET becomes the arbiter of truth for system behavior.
4. From FMEA Typist to Agentic AI Orchestrator
For Quality System Engineers, the shift is from manually populating compliance documents to governing systemic risk.
AI can help draft hazard analyses, traceability matrices, and regulatory documentation. But it cannot be blindly trusted to determine patient safety. The QSE’s role becomes even more critical as the human auditor of risk, logic, and clinical reality.
The key shift:
Govern the risk, don’t type the hazard.
This reframes quality work around judgment rather than documentation volume. The goal is not to produce more pages of compliance paperwork. The goal is to ensure that risk analysis is clinically meaningful, traceable, and logically sound.
5. From Director to Agentic AI Orchestrator
The final role shift may be the most important: Agentic AI transformation will fail if leadership remains trapped in IC Brain.
Leaders often ask why their teams are resisting AI. But if leaders continue to measure productivity by lines of code, story points, test scripts, FMEA rows, or pages of documentation, they are actively rewarding the old behavior.
The leadership shift is not simply to buy AI tools. It is to redesign the operating model.
The key shift:
Stop managing the assembly line and start architecting the factory.
Leaders must change the scoreboard, fund upskilling, redesign cross-functional loops, and give teams permission to become orchestrators rather than artifact generators.
The full Orthogonal Playbook goes deeper into each role shift, with tactical guidance for engineering, product, SDET, QSE, and leadership teams adopting Agentic AI in regulated environments.
Ready to move from AI experimentation to operating-model advantage?
Why This Matters for MedTech Teams
Agentic AI does not remove the need for expert people. It changes where expert people create value.
The engineer’s value moves toward architecture, constraints, security, and review. The Product Owner’s value moves toward clinical need, user value, and strategic prioritization. The SDET’s value moves toward behavioral contracts and verification integrity. The QSE’s value moves toward systemic risk and patient safety. The leader’s value moves toward organizational design and commercial outcomes.
In MedTech, where there is no margin for error, human judgment becomes more important in an agentic environment, not less.
The organizations that succeed with Agentic AI will not simply be the ones that deploy the most tools. They will be the ones that build a reviewer-first culture, write better constraints, encode richer clinical context, train people to govern digital labor, and align measurement systems around business and patient outcomes.
The Future Belongs to the Amplifiers
The goal of Agentic AI is not to make people type faster.
The goal is to give human experts more capacity to solve the problems only they can solve.
The agents are waiting for constraints. They are waiting for behavioral contracts, regulatory boundaries, and clinical context. And in many organizations, teams are waiting for permission to work differently.
Read Orthogonal’s playbook for engineers, Product Owners, SDETs, Quality System Engineers, and leaders.
Download Your Copy of the Playbook
Related Posts
Article
MedTech Teams Should Stop Paying for Compliance Twice
Article
Beyond the Device: Where MedTech Digital Ecosystems Are Starting to Pay Off
Article
Why MedTech Teams Slow Down as Software Grows and What Better Architecture Looks Like
