Extensibility & Platform Strategy

Building the platform other teams build on

Master APIs, SDKs, and plugin systems; the platform-product boundary; cross-product leverage; velocity versus platform health; GA readiness; multi-quarter roadmapping; and B2A API design.

Modules in This Level

7.9 B2A — Agents as API Consumers

Understand what changes when AI agents become Make's primary API consumers — machine-readable error codes, deterministic behavior, explicit capability declarations — and the B2A design principles that prevent agents from generating error traffic through trial-and-error discovery.

7.1 Extensibility as Platform Strategy

Understand why extensibility is a strategic posture rather than a feature — Figma's plugin ecosystem demonstrates the self-reinforcing flywheel of use cases, users, and builders — and identify the five extensibility surfaces Make must build before a competitor does it first.

7.2 APIs, SDKs, and Plugin Systems

Evaluate REST APIs, JavaScript SDKs, and plugin systems as distinct extensibility primitives — different audiences, maintenance costs, capability ceilings — and which to build first based on whether consumers are server integrations, JS developers, or marketplace partners.

7.3 The Platform-Product Boundary

Navigate the Supabase-specific versus generic-database-layer decision using variation confidence, extension potential, and team capacity — and learn when the abstraction argument is strongest versus weakest to resolve boundary debates without deferring to engineering instinct.

7.4 Cross-Product Platform Leverage

Evaluate Make's sandbox, preview infrastructure, and component runtime as shared services for Figma Design, FigJam, and Sites — and identify when the governance cost of serving multiple product teams (conflicts, prioritization disputes, velocity tax) exceeds the ROI.

7.5 Velocity vs. Platform Health

Replace 'the runtime needs a rewrite' with quantified risk analysis — load threshold, failure mode, user impact, probability — and evaluate three sequencing options (feature freeze, parallel track, strangler fig) so decisions use data, not competing intuitions.

7.6 GA Readiness — From Beta to Production

Learn the four GA readiness dimensions — 30 consecutive days meeting SLOs, 3x load testing with graceful degradation at 4x, penetration test and SOC 2 Type I, and runbooks any on-call engineer can use to resolve a P0 in under 30 minutes — so GA is a commitment, not a milestone.

7.7 Multi-Quarter Platform Roadmapping

Learn to sequence seven competing platform investments — runtime rewrite, preview upgrade, agent infrastructure, eval system, observability, GA hardening, and extensibility APIs — by mapping hard dependencies and the rule: harden what you've built, not what you plan to replace.

7.8 Level 7 Checkpoint

Apply the full course to a CPO-level scenario: given a 6-month GA deadline and plugin ecosystem launch 6 months later, present a phased platform strategy that sequences runtime hardening, observability, extensibility APIs, and developer trust in the order dependencies require.