Rigel Rise Research · April 2026

AI didn't add a phase to the development cycle. It broke the balance between them.

The SDLC was designed for humans. AI entered as an accelerator, not a redesign. Development exploded. Testing, validation, deploy, monitoring stayed roughly the same. The imbalance generates entropy that compounds with every cycle.

The symbiotic SDLC

Six phases. AI changed them unevenly. 41% of code is now AI-generated, but delivery throughput hasn't improved. A gap opens between what the system becomes and what humans understand about it. Each gap generates entropy. Each gap is where a product operates.

01 ANALYSIS02 DESIGN03 DEVELOPMENT41% of code is AI-generatedStackOverflow 202555% faster at writing codePeng et al. 202304 QA05 DEPLOYthroughput: -1.5%DORA 202406 MONITORINGstability: -7.2%DORA 2024the gapmore code, same deliveryREADYSIGNALDevelopers believe they are20% faster. Delivery datasays otherwise.METR 2025 · Faros AI 2025
Analysis
AI drafts specs and estimates faster. Commitment happens before understanding.
Design
AI proposes architectures with less deliberation. Trade-offs go unexamined.
Development
41% of code is AI-generated. Volume explodes. Understanding doesn't follow.
QA / Testing
More code to validate. Same infrastructure. Throughput drops 1.5%.
ReadySignal
Deploy
More changes to ship. Environments don't scale. Queues grow.
ReadySignal
Monitoring
Delivery stability drops 7.2%. Bugs increase. Ownership blurs.

75% of developers use AI. Companies see no measurable performance gains. Developers write more code, but organizations don't deliver faster. The gap between production speed and delivery capacity is where entropy accumulates. Each cycle, the gap widens.

Sources:41% AI-generated code Index.dev/StackOverflow 2025 · 55% faster Peng et al., GitHub/Accenture 2023 · Throughput -1.5%, Stability -7.2% DORA Report 2024 · 20% perceived, 19% slower METR 2025 · 75% adoption, no gains Faros AI 2025

Formal models

The gap per phase can be modeled. Two functions describe how systems enter cognitive failure and how entropy compounds across phase boundaries.

Strategic Equilibrium
(wR · Reflection) × (wC · Clarity)
(wU · Urgency) + (wCh · Chaos)
R
Revision before action. Measured by review frequency and cycle stability.
C
Shared understanding. Decreases when code outgrows documentation.
U
Pressure to act without analysis. Grows with deadlines and crises.
Ch
Disorganization. Technical (bugs, rework) or human (noise, conflicts).
Organizational Entropy
H = w1D + w2R + w3C
D
Dependencies between components, teams, and SDLC phases.
R
Unvalidated assumptions that accumulate silently across phases.
C
Information that doesn't reach the people who need it.

Entropy from interdependence and noise, not from lack of structure.

Why products compose

Each product covers one phase of the SDLC. When two cover neighboring phases, what each observes gains context from the other. Combined value is not additive. It multiplies.

READYSIGNALtested, validated, deployedUC-01UC-02UC-03Process paymentvalidated ✓Generate invoicevalidated ✓Update billingvalidated ✓CHECKOUT REDESIGN — DEPLOYEDUNVEILSIGNALbusiness meaning of each use caseUC-01UC-02UC-03Card authorizationvia Stripe gatewayTax calculationfor 12 countriesSyncs ledger withaccounting systemCHECKOUT FLOW — 2,000 CUSTOMERS+COMPOSEDoperational + business in one readingUC-01Process paymentStripe gateway · card authorizationvalidated + deployed ✓UC-02Generate invoiceTax calc · PDF · 12 countriesvalidated + deployed ✓UC-03Update billingLedger sync · accounting systemvalidated + deployed ✓"Checkout redesign deployed. Payment, invoicing, and billingvalidated across Stripe, tax engine, and ledger for 2,000 customers."
Operational
Business
Composed

Each agent is designed in isolation with defined inputs and outputs. The composition protocol makes them interoperable. The architecture follows the structure of the SDLC itself.

Where this goes next

Rigel Rise is the technical foundation. From here, two paths: see how these forces accumulate into business cost, or apply the ideas to validate delivery in your own cycle.

Papers & methodology

RR-001Software Entropy: Why Systems Degrade and What Controls the RateFoundationRR-002The Strategic Equilibrium Function: Decision Quality Under PressureMethodRR-003Composability as Entropy Resistance: Architecture for Cumulative ValueArchitectureRR-004The Symbiotic SDLC: How AI Broke the Phase BalanceThesisRR-005Organizational Entropy and Inter-Phase DependenciesModelRR-006Cognitive Maturity Curves and the Cost of Reactive Decision-MakingMethod