GUIDE Framework

Governed Use of Intelligent Design and Engineering

AI Leverage Operating Model

Version 4, Updated July 14, 2025

Purpose

GUIDE exists to restructure the software development process for an AI assisted environment.

As AI lowers the cost of design and engineering, the most important question is no longer how fast teams can build, but how teams restructure the development process to use AI effectively.

GUIDE shifts exploratory and validation work earlier so production code is touched less often and engineering effort is applied with greater precision.

AI increases total production capacity across design and engineering. GUIDE governs how that expanded capacity is directed, reviewed, and committed so increased output does not increase risk.

Why GUIDE Exists

For years, meaningful prototyping was slow and expensive. Validation often happened directly in production code.

Production code is an expensive place to learn.

When learning happens in production, engineers resolve uncertainty while managing architecture, integrations, security, and correctness. This increases cognitive load and downstream cost.

AI changes what is possible upstream. Teams can now explore workflows, interfaces, and edge cases through interaction before engineering effort is fully engaged.

GUIDE realigns development around this shift by separating exploratory work from production implementation.

What GUIDE Is

GUIDE is an operating framework for structuring software development in an AI assisted environment.

It makes intent, behavior, and feasibility concrete early so engineering effort is applied against clear direction rather than interpretation.

GUIDE ensures increased AI output does not increase system risk.

Core Heuristic

Show behavior, not representations.

Documents and diagrams describe systems. Interactive prototypes allow stakeholders to experience them.

GUIDE shifts validation into interactive, disposable prototypes that behave like finished products without production cost.

Disposable Means

• Not connected to production systems
• Safe to change or abandon
• Free from long term infrastructure commitment

Validation Before Commitment

Engineering effort is applied only after validation.

Work Is Ready for Engineering When:

1. A working prototype demonstrates core workflows and interactions
2. Stakeholders agree the prototype reflects what should be built
3. Edge cases and constraints are identified and documented
4. Open questions are explicit

Engineering begins when shared understanding exists.

Core Accountabilities

GUIDE concentrates decision accountability to reduce ambiguity and rework.

Design Owner

The Design Owner resolves uncertainty before engineering begins.

• Clarifies the problem and desired outcomes
• Produces a working, testable prototype
• Validates with stakeholders until alignment exists
• Surfaces edge cases and failure modes
• Governs AI usage in design artifacts

The prototype serves as the primary reference artifact for engineering.

Engineering Owner

The Engineering Owner governs feasibility and system integrity.

• Defines and enforces technical constraints
• Governs architecture and system boundaries
• Reviews validated design intent for feasibility
• Makes decisions regarding security, performance, and maintainability
• Governs AI usage in implementation

The Paired Control Loop

Design Owners validate direction.
Engineering Owners authorize build.

Client facing validation is reviewed for feasibility before presentation. Production implementation is authorized only after validated intent.

Governed Use of AI

AI is applied wherever it meaningfully reduces effort, but ownership remains human.

• AI may assist with prototypes, documentation, and implementation
• All output is reviewed against architectural and quality standards
• No artifact enters validation or production without explicit authorization

AI performs mechanical generation. Humans perform evaluative judgment and final commitment.

What GUIDE Does Not Prescribe

GUIDE does not define organizational structure or delivery rituals.

• Team size or staffing models
• Delivery cadence or sprint structure
• Estimation or velocity targets
• Roadmapping frameworks

These decisions remain with the organization.