PRISM D1 Velocity Assessment

AI-Assisted Development Lifecycle Maturity Report
Customer
NovaPay
Team Size
6 engineers
Funding Stage
Series A
Assessment Date
2026-04-10
Solutions Architect
Jordan Mitchell
Repository
github.com/novapay/nova-core

Executive Summary

L1.5
PRISM D1 Level 1.5
Experimental — Ad hoc AI use, no shared standards
Needs Foundations

NovaPay is a Series A startup with 6 engineers, currently assessed at PRISM D1 Level 1.5 (Experimental). The automated scanner scored 19/100 and the SA interview scored 28/100, with an org readiness score of 12/20, producing a blended score of 30.8. The assessment verdict is Needs Foundations, routing NovaPay to Track A: Foundations.

19
/ 100
Scanner Score (40%)
28
/ 100
Interview Score (40%)
12
/ 20
Org Readiness (20%)

Scanner Category Breakdown

AI Tool Config Spec-Driven .. Commit Hygiene CI/CD Integr.. Eval & Quality Testing Matu.. AI Observabi.. Governance Agent Workfl.. Platform Reuse Documentation Dependencies
CategoryScoreProgressStatus
AI Tool Config 3 / 10
30%
Spec-Driven Dev 1 / 10
10%
Commit Hygiene 4 / 15
26%
CI/CD Integration 5 / 15
33%
Eval & Quality 0 / 10
0%
Testing Maturity 3 / 10
30%
AI Observability 0 / 10
0%
Governance 0 / 5
0%
Agent Workflows 0 / 5
0%
Platform Reuse 1 / 5
20%
Documentation 1 / 3
33%
Dependencies 1 / 2
50%

Interview Assessment

SectionScoreStatus
AI Tooling Landscape
  • Three developers use Claude Code but with personal configurations
  • No CLAUDE.md or shared AI tool configuration
  • Two developers still using ChatGPT copy-paste workflow
  • No Kiro IDE adoption
 6 / 15 40%
Development Workflow & Specs
  • No spec-driven development practice in place
  • AI prompts are ad-hoc, not documented or reused
  • Feature requirements exist in Jira but are not formatted as AI specs
  • Team lead expressed strong interest in structured AI workflows
 3 / 20 15%
CI/CD & Quality
  • GitHub Actions pipeline exists with basic lint and test steps
  • No eval gates or AI-specific quality checks
  • Test coverage at 42%, below industry average
  • No differentiation between AI-generated and human-written code in CI
 8 / 20 40%
Metrics & Visibility
  • No AI-specific metrics tracked at all
  • No dashboard or executive visibility into AI usage
  • Cannot distinguish AI-assisted commits from manual ones
  • Team lead estimates 30-40% of code is AI-assisted but has no data
 1 / 15 6%
Governance & Security
  • No formal AI usage policy
  • No data handling guidelines for AI tools
  • PII occasionally passed to AI tools without awareness
  • No approval process for new AI tool adoption
 2 / 15 13%
Org & Culture
  • CTO is enthusiastic about AI-assisted development
  • Most developers are eager to adopt structured AI workflows
  • One developer is resistant but open to evidence-based persuasion
  • Small team size makes adoption changes easier
 8 / 15 53%

Organizational Readiness

✓Executive Sponsor Identified
✗Dedicated AI Champion
✓Team Willingness
✗Budget Approved
✓Timeline Commitment

Org readiness score: 12/20

Top Strengths

  1. Dependencies: 1/2 (50%)
  2. CI/CD Integration: 5/15 (33%)
  3. Documentation: 1/3 (33%)

Gap Analysis & Remediation

PriorityCategoryScoreRecommended Action
#1 Eval & Quality (scanner) 0/10 (0%) Define quality rubrics for AI-generated code. Implement automated scoring. Establish minimum acceptance thresholds.
#2 AI Observability (scanner) 0/10 (0%) Deploy the EventBridge-Timestream metrics pipeline. Enable token tracking, cost attribution, and latency monitoring.
#3 Governance (scanner) 0/5 (0%) Create an AI usage governance charter. Define approval workflows for agent autonomy levels. Document data handling policies.

Onboarding Recommendation

Track A: Foundations

Workshop Modules

IDModuleRationale
✓ M00 Environment Setup Foundation for everything else
✓ M01 CLAUDE.md & Standards Standardize AI tool configuration
✓ M02 Spec-Driven Development Introduce structured AI workflows
✗ M03 CI/CD & Eval Gates Too early -- no baseline to gate against
✗ M04 Metrics & Dashboards Too early -- need data flowing first
✗ M05 Governance & Scaling Too early -- need adoption first

Success Metrics

MetricTargetMeasure By
AI-origin commit trailers 30%+ of commits 2 weeks post-workshop
CLAUDE.md deployment 100% of active repos 1 week post-workshop
Spec adoption 2+ specs per developer 2 weeks post-workshop
Team tool satisfaction 7+ NPS score 2 weeks post-workshop

90-Day Roadmap

TimelineMilestoneMeasurable Outcome
Week 1 CLAUDE.md deployed, hooks active 100% of repos have CLAUDE.md
Week 2 First spec-driven features completed 2+ features built via spec workflow
Week 3 Commit trailer adoption steady 30%+ of commits have AI-origin trailers
Week 4 Re-assessment complete Scanner re-run, L2.0+ achieved for Track B upgrade
Week 8 Track B started or extended coaching Full workshop scheduled or coaching plan in place
Week 12 L2.5+ achieved Blended assessment score confirms L2.5+

SA Engagement Cadence

WeekTypeDurationAgenda
Week -2 Pre-work kick-off call 30 min Walk through pre-work checklist, answer setup questions, confirm workshop date
Week 0 Workshop delivery 4 hr Deliver Modules 00-02, hands-on exercises, first spec-driven build
Week 1 Async check-in 15 min (Slack) Review commit trailer adoption rates, troubleshoot blockers
Week 2 Video call 30 min Review first 2-week metrics, adjust approach if needed
Week 4 Re-assessment session 1 hr Re-run scanner, conduct brief interview, determine Track B readiness

PRISM D1 Velocity Assessment Report — Generated 2026-04-10

AWS Solutions Architecture · Startups Organization · CONFIDENTIAL

Solutions Architect: Jordan Mitchell