AI Project Lifecycle · Optional Service Module

Training & Change Management
for AI Projects.

AI adoption is not a technology implementation problem.

It is a judgment development problem.

Organizations do not fail at AI because their software is wrong. They fail because their people do not know how to govern probabilistic outputs — when to trust the system, when to override it, and who is accountable when it is wrong. RATIO's training and change management module builds that capability deliberately, integrated into the project governance lifecycle rather than bolted on at the end.

Build Judgment · Not Just Familiarity · Govern Outputs · Sustain Capability

AI change management is not
IT change management.

Traditional change management assumes you are implementing a known system with known outcomes. AI is different in kind, not just degree. Here is what that means for how you prepare your people.

Traditional IT Change Management
Learn the new software.
Follow the new process.

The system produces deterministic outputs. Right answers exist. Training covers how to use the tool correctly. Change management covers communication, adoption, and resistance. Once deployed, the system is stable.

AI Change Management
Develop judgment about
probabilistic outputs.

The system produces probabilistic outputs. Right answers are distributions, not certainties. Training covers when to trust, when to override, and how to document the decision. Change management covers accountability design. And the system drifts -- so the work is never done.

Four reasons standard training
fails in AI environments.

01
Working alongside distributions requires a different cognitive model than using software.

People trained on deterministic systems develop habits of deference -- if the system says it, it must be right. AI systems require active skepticism, structured review, and documented override decisions. That is a different kind of training entirely.

02
Organizational error compounds the same way agent error compounds.

If every reviewer applies the same uncalibrated judgment, the organizational error rate mirrors the agent error rate. Training that builds identical responses across a team creates the same homogeneity risk as buying the same framework as your competitors.

03
Training should build judgment, not conformity.

If everyone in your organization is trained the same way on the same frameworks, you recreate the herd problem internally. Practitioners need shared vocabulary and shared standards -- but their applied judgment must reflect your specific context, data, and risk environment.

04
AI change management is never done.

Model drift means human calibration must be continuous, not a one-time event at go-live. Business conditions change. Policies change. Competitor behavior changes. The training completed at deployment is not sufficient for year two. Organizational capability must evolve alongside the system.

Three insertion points.
Not one big training event.

Training and change management activate at the right moments in the project governance cycle -- before deployment, at quality gates, and continuously as the system and organization evolve.

Phase 1
Pre-Deployment
Readiness

Building practitioner capability before the system goes live. What do people need to understand about probabilistic outputs, the 5C framework, and when to trust versus override? This is where the foundational curriculum activates.

Covers
AI output literacy · Confabulation recognition · OCC-5C review protocol · Role and accountability orientation
Phase 2
Quality Gate
Preparation

Role-specific readiness at each governance decision point. The approval agent reviewer needs different preparation than the board-level governance authority. Just-in-time, tied directly to the accountability chain, not generic curriculum.

Covers
Gate-specific review criteria · WoE verdict authority · Escalation protocols · Documentation standards
Phase 3
Continuous
Calibration

As the model drifts, as business conditions change, as new practitioners join. Human calibration must evolve alongside the system. Periodic recalibration sessions tied to model retraining cycles and organizational change events.

Covers
Drift detection awareness · Retraining triggers · New practitioner onboarding · Competency reassessment

Three formats.
One coherent framework.

Every organization's readiness gap is different. Delivery format is matched to your context, scale, and culture -- not a one-size-fits-all curriculum.

Format 01
Instructor-Led
Workshops

Facilitated sessions customized to your organization's AI context, data environment, and risk profile. Designed for leadership teams, governance authorities, and cross-functional practitioners who need shared vocabulary and applied judgment -- not generic AI awareness training.

Best for
Leadership alignment · Governance authority preparation · High-stakes deployment readiness
Format 02
Asynchronous
Video Curriculum

Scalable, self-paced modules deployable across large practitioner populations. Competency-aligned to the OCC framework -- AI users, AI reviewers, and governance authorities each follow a distinct track. Built for enterprise scale without proportional labor cost.

Best for
Large practitioner populations · Distributed teams · Onboarding new hires into AI governance roles
Format 03 · Distinctive
AI Guy™
Guided Learning

Short animated modules using the AI Guy™ character as guide -- the same accessible, concept-first format as the Innovation With Trust editorial cartoon series. Each module covers a practitioner competency through a real scenario, a calculation or framework, and a principle to apply. Conceptual, memorable, and on-brand for organizations that want something different from a slide deck.

Best for
Organizations that want practitioner development to feel like insight rather than compliance · Culture-first AI adoption

Curriculum design anchored in the OCC framework.

Every training module maps to a specific competency in the OCC Organizational Cognitive Competence matrix -- the only framework that differentiates what AI users, AI reviewers, and AI governance authorities each need to know and demonstrate.

This means training is not generic AI awareness. It is role-specific, outcome-measurable, and tied directly to the accountability chain in your governance architecture. People leave knowing what their role is, what they are authorized to decide, and how to document it.

See the Full Competency Validation Service →
Branch A · AI Users
Confabulation recognition. Output review. Scope-of-validity assessment.
Knowing when the output is plausible but wrong -- and having the language to say so.
Branch B · AI Reviewers
Gestalt audit protocol. Hybrid cognition. WoE verdict authority.
Structured review of AI outputs against evidence standards. Issuing and documenting verdicts.
Branch C · Governance Authority
Accountability chain design. FATAL escalation. Board-level risk translation.
Owning the decision to proceed or stop. Translating AI risk into organizational consequence.

Every organization's AI readiness gap is different.

"Start with the human system.
Then decide where AI belongs."

Whether you are preparing for a first deployment, rethinking a governance structure that is not working, or building long-term practitioner capability, let's talk about what your organization actually needs.

Talk to John → Talk to Dale →
Also available as part of the AI Innovation With Trust Program · Learn more →