- 01 Figure , replaces Discovery. Signed: Aspiration, personas, surfaces, prototype.
- 02 Frame , replaces Design. Signed: Architecture, data model, governance, acceptance criteria.
- 03 Forge , replaces Build and QA. Signed: Plan, scored builds.
- 04 Field , replaces Support. Signed: Drift, compliance and cost reports.
Intelligence Studio
Our AI development lifecycle: four phases, named agents, and a human signature at every gate.
Why a new lifecycle
Traditional delivery was designed for people writing every line. When AI generates code at volume, reviews become the bottleneck, hidden data complexity surfaces mid-build, and business stakeholders still see the result only at user acceptance testing. Speed without a method creates rework.
Perspective · AIDLC P01 Read the paper: the SDLC was built for humans
The four phases
foundation
01 Figure
- Replaces
- Discovery
- What happens
- Intent becomes a signed contract. Where a system already exists, it is read and its intent reconstructed. A prototype is approved before any code is written.
- Signed output
- Aspiration, personas, surfaces, prototype
02 Frame
- Replaces
- Design
- What happens
- Architecture becomes the box the AI must build inside. Your architects and risk lead sign it first.
- Signed output
- Architecture, data model, governance, acceptance criteria
03 Forge
- Replaces
- Build and QA
- What happens
- Code is generated only against approved artifacts. Every sprint is scored against ground truth, and anything that fails is refined before it advances.
- Signed output
- Plan, scored builds
04 Field
- Replaces
- Support
- What happens
- Production is watched for schema and pattern drift. Compliance and cost are reviewed continuously.
- Signed output
- Drift, compliance and cost reports
Four principles
- Decisions before code
- Architecture, data and governance are signed before code is generated.
- Business sees it first
- A working prototype is approved before the build.
- Every output is scored
- Acceptance criteria with ground truth, evaluated every sprint.
- It transfers
- The method lives as plain-text artifacts in your repository.
What you keep
- The code. Every line lives in your repository from the first sprint.
- The artifacts. Architecture, data model, governance posture, acceptance criteria and decision records.
- No runtime dependency. No key we hold, no service you rent.
- Off-ramps at every gate. Stop after Figure and you keep the discovery. Stop after Frame and you keep the architecture.
Toolkits
-
MCP Server Pack
Azure-hosted MCP scaffolds for enterprise data, documents and signals
-
Cowork Workflow Pack: Front Office
Briefing, drafting and research workflows for front-office teams
-
Agentic Guardrails & Prompts
Contracts and modules aligned to each Intelligence Studio phase
-
AI Readiness & Security Assessment
Estate, data and security readiness scored against a fixed rubric
-
Adoption Playbook
Enablement, guardrails and change management, ready to run
Every Scorpiosys engineer is certified by delivering under Intelligence Studio
Inside the Academy