The NUP Lifecycle for Probabilistic Software

Last Audited: 2026-08-24
NUP AI-Native Verified
ISO/IEC 42001:2023 Cl. 6, 8, 9NIST AI RMF 1.0 Govern, Map, Measure, ManageEU AI Act Art. 9, 10, 14, 15, 61
In Plain Language

Traditional software development lifecycles (such as Waterfall or Agile Scrum) rely on binary phase gates: write code, pass 100% unit tests, sign off UAT, and deploy. In probabilistic systems, where outputs vary validly and models drift in production, applying deterministic milestones leads to project stalls. The NUP Probabilistic Lifecycle establishes a 6-stage engineering framework—from Discovery and Knowledge Prep to AI Development, Statistical Validation, Staged Release, and Continuous Drift Monitoring—with concrete audit deliverables mapped to ISO 42001, EU AI Act, and NIST AI RMF.

Architectural Orientation: Why AI Demands a New Lifecycle

Engineering leaders cannot manage AI initiatives using the assumption that software artifacts remain frozen post-release. In AI-native architectures, enterprise intelligence evolves continuously: training data shifts, vector indexes expand, model providers update foundation weights, and users prompt models in unforeseen ways.

The Netspective Unified Process (NUP) for Probabilistic Software defines a six-stage lifecycle that replaces binary sign-offs with continuous statistical bounds, feedback loops, and tamper-evident audit evidence.

The 6-Stage NUP Probabilistic Lifecycle ArchitectureAn overview diagram showing the six sequential stages of the Netspective Unified Process for Probabilistic Software, with continuous feedback loops and governance boundaries.The 6-Stage NUP Probabilistic LifecycleSequential audit-ready phase gates paired with continuous feedback loops and post-market governance.CONTINUOUS GOVERNANCE & EVIDENCE ACCUMULATION (ISO 42001 / EU AI ACT / NIST AI RMF)STAGE 01Discovery & Framing• AI Applicability• Bias Assessment• Regulatory Tiers• Success MetricsArtifact: Intent SpecSTAGE 02Knowledge & Data• Trust Layers• Chunk Lineage• Context Strategy• Dataset VersioningArtifact: Data SheetSTAGE 03AI-Augmented Dev• Prompt & Context• Harness & Loops• Human Oversight• Code ProvenanceArtifact: Code LogSTAGE 04Validation & Test• Statistical Eval• Hallucination %• Red Teaming• Citation GroundingArtifact: Eval ReportSTAGE 05Deploy & Release• Shadow Execution• Canary Rollouts• Human Escalation• AI DisclosuresArtifact: Rollout LogSTAGE 06Monitor & Improve• Drift Telemetry• Feedback Loops• Retraining Cycles• Incident LogsArtifact: Drift ReportContinuous Feedback Loop: Production Telemetry & User Corrections ➔ Knowledge Base Refresh (Stage 2)

Interactive Phase Explorer

Click through each stage to explore its primary activities, risks, exit criteria, and compliance artifacts:

Interactive SDLC Inspector

NUP 6-Stage Probabilistic Lifecycle Explorer

Use arrow keys or click tabs to explore
STAGE 1 OF 6

Discovery & Problem Framing

Defines whether the problem requires probabilistic AI or traditional deterministic algorithms. Assesses data quality, regulatory exposure (FDA SaMD, EU AI Act risk tiers), and establishes statistical success metrics.

Primary Engineering Activities
  • Evaluate feasibility: Probabilistic vs. Deterministic problem categorization
  • Conduct initial AI risk classification (EU AI Act Prohibited / High-Risk / Minimal)
  • Assess training/context data availability, copyright provenance, and representation bias
  • Define non-deterministic success criteria (acceptable error rates, latency budgets, cost per token)
  • Plan knowledge transformation and data extraction requirements from source documents
Probabilistic Failure Risks
  • Applying LLMs to problems requiring 100% exact math or deterministic business rules
  • Underestimating latency and token consumption at enterprise scale
  • Failing to identify regulatory high-risk classification before architecture freeze
Stage Exit Criteria
  • Approved Problem Framing Brief with explicit probabilistic justification
  • Initial AI Risk Classification Matrix signed off by compliance officer
  • Target statistical metrics defined (Target Accuracy ≥ 95%, Hallucination Rate ≤ 1%)
Auditable Artifacts
AI System Intent Specification (ISO 42001 Cl. 6.1)Preliminary Risk & Impact Assessment (EU AI Act Art. 9)Knowledge Sourcing & Provenance Manifest
Regulatory Standards
NIST AI RMF 1.0 · GOVERN 1.1, MAP 1.1EU AI Act · Article 6 (Risk Classification)ISO/IEC 42001:2023 · Clause 6.1 (Actions to address risks)
STAGE 01 OF 06

Discovery & Problem Framing

Determining applicability, bias risks, and non-deterministic success criteria

Defines whether the problem requires probabilistic AI or traditional deterministic algorithms. Assesses data quality, regulatory exposure (FDA SaMD, EU AI Act risk tiers), and establishes statistical success metrics.

Key Engineering Activities

  • Evaluate feasibility: Probabilistic vs. Deterministic problem categorization
  • Conduct initial AI risk classification (EU AI Act Prohibited / High-Risk / Minimal)
  • Assess training/context data availability, copyright provenance, and representation bias
  • Define non-deterministic success criteria (acceptable error rates, latency budgets, cost per token)
  • Plan knowledge transformation and data extraction requirements from source documents
Probabilistic Risks & Anti-Patterns
  • Applying LLMs to problems requiring 100% exact math or deterministic business rules
  • Underestimating latency and token consumption at enterprise scale
  • Failing to identify regulatory high-risk classification before architecture freeze

Stage Exit Criteria

Approved Problem Framing Brief with explicit probabilistic justification
Initial AI Risk Classification Matrix signed off by compliance officer
Target statistical metrics defined (Target Accuracy ≥ 95%, Hallucination Rate ≤ 1%)
Audit Evidence Artifacts & Regulatory Mappings
Mandatory Audit Deliverables:
AI System Intent Specification (ISO 42001 Cl. 6.1)
Preliminary Risk & Impact Assessment (EU AI Act Art. 9)
Knowledge Sourcing & Provenance Manifest
Governing Compliance Clauses:
NIST AI RMF 1.0 GOVERN 1.1, MAP 1.1EU AI Act Article 6 (Risk Classification)ISO/IEC 42001:2023 Clause 6.1 (Actions to address risks)
STAGE 02 OF 06

Knowledge & Data Preparation

Transforming enterprise assets into provenance-tracked, AI-consumable context

Converts unstructured enterprise documents (PDF, Word, Excel) into structured, metadata-rich formats (Markdown, HTML) for retrieval-augmented generation. Analyzes data distributions and establishes strict lineage.

Key Engineering Activities

  • Transform complex legacy documents into clean, chunkable Markdown and HTML trust layers
  • Establish cryptographic data provenance and chunk-to-source traceability
  • Analyze datasets for demographic imbalances, sampling bias, and toxic content
  • Design context engineering strategies (chunk sizing, metadata tagging, hybrid sparse/dense indexing)
  • Version and snapshot knowledge bases and vector indexes in tamper-evident registries
Probabilistic Risks & Anti-Patterns
  • Garbage-in, garbage-out: Poorly OCR’d or unformatted tables corrupting LLM context
  • Stale vector embeddings producing hallucinated answers based on deprecated policies
  • Data leakage: Embedding proprietary or PII data without role-based access control filters

Stage Exit Criteria

100% of knowledge base chunks mapped to authoritative source documents with citations
Data Sheet for Datasets completed and verified
Vector and lexical search retrieval recall score ≥ 90% on benchmark queries
Audit Evidence Artifacts & Regulatory Mappings
Mandatory Audit Deliverables:
Data Sheet for Knowledge Base (NIST AI RMF MEASURE 2.1)
Chunking & Context Engineering Specification
Data Lineage & Provenance Log (ISO 42001 Cl. 8.2)
Governing Compliance Clauses:
EU AI Act Article 10 (Data & Data Governance)FDA AI/ML GMLP Principle 3 (Data representativeness)NIST AI RMF 1.0 MAP 2.1, MEASURE 2.2
STAGE 03 OF 06

AI-Augmented Development

Orchestrating prompt, context, harness, and loop engineering with human oversight

Executes implementation using the 4 Layers of LLM Engineering (L1 Prompt, L2 Context, L3 Harness, L4 Loop). Employs AI coding assistants with strict code provenance and human-in-the-loop review.

Key Engineering Activities

  • Implement prompt templates with explicit instructions, constraints, and few-shot exemplars
  • Build context injection middleware (surveilr SQL-native middleware, dynamic RAG pipelines)
  • Implement harness engineering: validation guards, output schema validators (JSON Schema/Zod)
  • Construct agentic loop engineering with bounded recursion, step budgets, and human interruption points
  • Track AI-generated code provenance and conduct peer reviews on all synthesized logic
Probabilistic Risks & Anti-Patterns
  • Prompt injection vulnerabilities allowing malicious user input to hijack system instructions
  • Unbounded agentic loops resulting in infinite tool execution and API cost spikes
  • Accumulation of AI-generated tech debt without adequate human architectural understanding

Stage Exit Criteria

All prompt templates cataloged in versioned registry with test fixtures
Output schema validation enforced on 100% of LLM tool and API responses
Human-in-the-loop review approved for all autonomous agent execution paths
Audit Evidence Artifacts & Regulatory Mappings
Mandatory Audit Deliverables:
Prompt & Context Architecture Specification
AI Tool Usage & Code Provenance Register (ISO 27001 A.8.25)
Guardrail & Schema Validation Policy
Governing Compliance Clauses:
EU AI Act Article 14 (Human Oversight)ISO/IEC 42001:2023 Clause 8.4 (AI system development)NIST AI RMF 1.0 MANAGE 2.1, MANAGE 2.3
STAGE 04 OF 06

Validation & Testing

Continuous statistical evaluation, distribution analysis, and adversarial red teaming

Replaces binary unit testing with statistical validation across sample populations. Conducts hallucination measurement, semantic drift tracking, bias audits, and red-team adversarial attacks.

Critical Lifecycle Distinction: Statistical vs. Deterministic Gate

Unlike deterministic testing which enforces binary exact matches, Stage 4 validation requires statistical confidence evaluation across sample populations (e.g. 95% confidence bounds on hallucination rates < 0.8% and retrieval recall ≥ 90%).

Deep Dive into Safety & Hallucination Mitigation (Trust & Retrieval Engineering)

Key Engineering Activities

  • Execute statistical evaluation suites across test populations with 95% confidence intervals
  • Measure hallucination rates using automated LLM-as-judge and ground-truth verification
  • Perform adversarial red teaming to probe for prompt extraction, jailbreaks, and harmful content
  • Conduct bias and fairness audits across protected demographic attributes
  • Verify citation correctness: ensure every generated claim links to an intact source chunk
Probabilistic Risks & Anti-Patterns
  • Sampling bias in evaluation benchmarks leading to false confidence in edge cases
  • Over-reliance on LLM-as-a-judge without human verification of judge reliability
  • Failing to test multi-turn conversation escalation and adversarial memory poisoning

Stage Exit Criteria

Statistical benchmark accuracy satisfies SLA across 1,000+ representative test cases
Red-team penetration test completed with zero critical jailbreak vulnerabilities open
Citation grounding accuracy verified at ≥ 98% on critical enterprise knowledge queries
Audit Evidence Artifacts & Regulatory Mappings
Mandatory Audit Deliverables:
Statistical Validation & Evaluation Report (NIST AI RMF MEASURE 1.1)
Red Team & Adversarial Assessment Summary
Bias & Fairness Audit Log (EU AI Act Art. 10)
Governing Compliance Clauses:
NIST AI RMF 1.0 MEASURE 1.1, MEASURE 2.6EU AI Act Article 15 (Accuracy, Robustness, Cybersecurity)FDA SaMD GMLP Principle 7 (Focus on total product life cycle)
STAGE 05 OF 06

Deployment & Release

Staged canary rollouts, human escalation gates, and transparent user disclosures

Implements progressive delivery mechanisms (shadow pipelines, canary rollouts) for probabilistic features. Enforces human-in-the-loop escalation gates for high-stakes actions and delivers clear AI disclosures.

Key Engineering Activities

  • Deploy AI features in shadow mode to compare model outputs against legacy human workflows
  • Execute canary rollouts (5% ➡️ 25% ➡️ 100%) with real-time anomaly tripwires
  • Establish human-in-the-loop escalation workflows for outputs with confidence scores < 0.85
  • Deploy transparent user-facing AI disclosures ("Generated by AI, verified by human")
  • Configure AI gateway routing, rate limiting, token quotas, and prompt caching layers
Probabilistic Risks & Anti-Patterns
  • Deploying AI systems directly to 100% of users without shadow comparison baseline
  • Lack of clear user disclosure violating EU AI Act transparency mandates
  • Sudden upstream model API deprecation or unannounced behavior modifications

Stage Exit Criteria

Shadow mode performance parity verified over minimum 14-day production observation window
Canary deployment passes all automated error budget and latency criteria
Human escalation routing tested and operational for customer support teams
Audit Evidence Artifacts & Regulatory Mappings
Mandatory Audit Deliverables:
Deployment & Rollout Verification Record
Predetermined Change Control Plan (FDA PCCP)
AI Transparency Disclosure & User Notice Specification
Governing Compliance Clauses:
EU AI Act Article 50 (Transparency Obligations)FDA SaMD PCCP Section IV (Modification protocol)ISO/IEC 42001:2023 Clause 8.5 (AI system operation)
STAGE 06 OF 06

Monitoring & Continuous Improvement

Real-time drift detection, feedback aggregation, and scheduled retraining cycles

Monitors output distributions, retrieval latency, and semantic drift in production. Closes the feedback loop by ingesting user corrections, updating knowledge bases, and executing scheduled evaluation sweeps.

Key Engineering Activities

  • Continuously stream and analyze production inference telemetry for semantic drift
  • Collect implicit and explicit user feedback (thumbs up/down, user prompt edits, copy actions)
  • Track model latency, token costs, cache hit rates, and error classifications in real time
  • Execute automated weekly evaluation sweeps against frozen golden benchmark datasets
  • Maintain an incident response playbook for prompt injection attacks and harmful outputs
Probabilistic Risks & Anti-Patterns
  • Silent semantic drift: Model outputs gradually degrading over time without error alerts
  • User feedback poisoning: Malicious users submitting false feedback to skew retraining data
  • Unmonitored token cost inflation during high-traffic enterprise usage spikes

Stage Exit Criteria

24/7 telemetry monitoring active with automated drift alerts connected to PagerDuty
Monthly Continuous Governance Review completed with updated risk matrices
Feedback aggregation pipeline feeding verified improvements back into Stage 2 knowledge bases
Audit Evidence Artifacts & Regulatory Mappings
Mandatory Audit Deliverables:
Continuous Monitoring & Drift Telemetry Report (ISO 42001 Cl. 9.1)
Safety Incident & Corrective Action Log (EU AI Act Art. 62)
Quarterly Model Performance & Retraining Audit
Governing Compliance Clauses:
ISO/IEC 42001:2023 Clause 9.1 (Monitoring, measurement, analysis)EU AI Act Article 61 (Post-Market Monitoring)NIST AI RMF 1.0 MANAGE 4.1, MANAGE 4.3
Try This with AI: 6-Stage Lifecycle Gate Audit Prompt

Use this prompt in your AI coding assistant to audit your current project state against NUP Probabilistic exit criteria.

Act as an AI Quality & Regulatory Compliance Auditor. Review the following project state against the Netspective Unified Process for Probabilistic Software: - Current Lifecycle Stage: [e.g., Stage 4: Validation & Testing] - System Architecture: [e.g., Multi-turn RAG chatbot utilizing GPT-4o with pgvector and surveilr middleware] - Completed Artifacts: [e.g., 500-sample benchmark eval dataset, Prompt Registry v1.2, Red team summary] Evaluate: 1. Does the project satisfy the Stage 4 exit criteria for statistical evaluation and hallucination measurement? 2. What specific evidence artifacts (e.g. EU AI Act Annex IV technical dossier, NIST AI RMF profile) are missing? 3. Provide a step-by-step checklist of tasks required before approving promotion to Stage 5 (Deployment & Release).
Next in Core Concepts

Topic 5: Evidence Requirements for Probabilistic Systems

Proceed to Topic 5
Previous
The AI-Native Gap
Core Concepts
Next
Evidence Requirements for Probabilistic Systems
Core Concepts

Community Discussion & Feedback

Attributed peer feedback and official Netspective architecture notes.

Was this documentation helpful?(100% found this helpful • 0 ratings)

Leave Feedback or Question

○ Loading user info...
0/2000 chars

Discussion (0)

Loading discussion thread...