Prioritization as code: An AI-supported framework for platform engineering

Engineering Enablement by DX

0:00

-38:50

Prioritization as code: An AI-supported framework for platform engineering

How SiriusXM built a data-driven framework for platform engineering prioritization by making assumptions visible, measurable, and easier to challenge.

Justin Reock

Jun 15, 2026

Listen and watch now on YouTube, Apple, and Spotify.

In this session from DX Annual, Eleanor Millman, Senior Staff Product Manager, and Mina Tawadrous, Associate Director of Product Management at SiriusXM, share how their platform engineering organization developed a prioritization framework for platform engineering teams serving hundreds of developers across a complex cloud platform.

They explain how they define and weight platform-specific impact factors, use developer data to refine priorities, and score projects more consistently. They also explore why prioritization debates often stem from conflicting, invisible, or outdated assumptions, and how SiriusXM began treating assumptions like code by documenting, versioning, and reviewing them in source control.

Finally, they demonstrate how AI can surface assumptions, connect initiatives to existing knowledge, and support project scoring while keeping humans in the loop. Throughout the session, they offer a practical framework for making prioritization decisions more transparent, data-driven, and scalable.

Some takeaways:

Building a platform engineering prioritization framework

Platform engineering requires different prioritization criteria. SiriusXM found that traditional product metrics did not fully capture the value of platform engineering work, leading the team to define platform-specific impact factors around development speed, reliability, security, cost, platform efficiency, user trust, and data-driven decision making.
A simple scoring model created a shared language for prioritization. The framework combined impact, urgency, effort, and business needs to help teams compare projects consistently and explain why certain initiatives were prioritized over others.
The framework evolved alongside the organization. As company priorities changed after a major platform launch, SiriusXM adjusted impact factor weights to reflect new goals around cost optimization, technical debt reduction, and data maturity.

Using developer data to guide decisions

Developer feedback helped shape prioritization. Rather than relying solely on intuition, the team used survey data and other developer insights to determine where additional investment would have the greatest impact.
Impact factor weights were revisited regularly. Quarterly reviews allowed the team to adjust priorities based on changing business objectives and improvements in areas such as reliability and security.
Data increased confidence in prioritization decisions. By grounding discussions in evidence, teams were able to align more effectively on where to invest their limited capacity.

Treating assumptions like code

Many prioritization conflicts stem from assumptions rather than priorities. Teams often disagreed because they were working from different, invisible, or outdated assumptions about users, workflows, and business needs.
Documenting assumptions improved organizational alignment. SiriusXM began storing assumptions in source control, making them easier to discover, review, update, and validate over time.
Debates became more productive when assumptions were explicit. Instead of arguing over which project mattered most, teams focused on validating the underlying beliefs that informed their decisions.

Using AI to surface organizational knowledge

Assumption repositories became difficult to navigate at scale. As more assumptions were documented, it became increasingly difficult for individuals to find relevant context and connections across projects.
AI helped uncover relationships humans might miss. By searching assumption repositories, OKRs, and prior project data, AI was able to surface relevant information that would otherwise be difficult to discover.
AI improved information recall rather than replacing judgment. The goal was not automated decision making but helping teams access the knowledge needed to make better decisions.

Building an AI-assisted prioritization workflow

AI can guide teams through the scoring process. SiriusXM built workflows that ask clarifying questions, surface assumptions, identify relevant organizational context, and generate initial project scores.
Human validation remains essential. Teams review assumptions, challenge recommendations, and approve updates before information is added back into the system.
Each prioritization cycle strengthens the knowledge base. New assumptions, decisions, and project context become available for future initiatives, making the system more valuable over time.

Keeping humans in the loop

The framework is designed to support conversations, not replace them. Scores help teams discuss priorities more objectively, but important decisions still require context and judgment.
Stakeholder disagreements often reveal useful information. When the framework produces results that feel wrong, the discussion can uncover missing assumptions, incomplete data, or opportunities to improve the model itself.
The framework continues to evolve. SiriusXM treats both the prioritization model and the supporting AI tools as products that require ongoing iteration, feedback, and refinement.