Implementing boundaries between expert advice, AI drafts, and client deliverables

Assessment finding

Professional judgments where AI must not decide are not explicit.

Service B implementation

Create Decision Boundaries by professional judgment type.

Assessment finding

Review triggers for client submission, legal impact, and financial impact are insufficient.

Service B implementation

Define review triggers and approvals before client submission.

Assessment finding

AI-specific review criteria are missing.

Service B implementation

Create a professional services Human Judgment Review procedure for sources, assumptions, and client conditions.

Assessment finding

Differences between AI drafts and final deliverables are not recorded.

Service B implementation

Introduce Decision Logs for AI drafts, references, human edits, and final deliverables.

Assessment finding

Boundary review is not tied to client contracts or confidentiality obligations.

Service B implementation

Define Boundary Governance for client conditions, new tools, and deliverable types.

Workflow

Research

AI use

Information gathering, summarization, issue candidates

Design focus

Require source, recency, and client-condition checks.

Workflow

Proposal creation

AI use

Structure, draft copy, comparison tables

Design focus

Define quality review, financial impact review, and feasibility confirmation before client submission.

Workflow

Contract and policy review

AI use

Issue candidates, clause summaries, risk candidates

Design focus

Limit AI to issue candidates; human experts retain legal and advisory judgment.

Workflow

Client-facing deliverable drafting

AI use

Drafting, tone adjustment, summarization

Design focus

Require expert review, client-condition checks, and Decision Log records.

Deliverable

Professional judgment Decision Boundary™

Description

Defines AI role and human judgment across advice, review, proposals, and deliverables.

Primary users

Engagement owners, quality management, professionals

Deliverable

Review trigger and threshold list

Description

Defines review conditions for client submission, legal impact, financial impact, and critical proposals.

Primary users

Engagement owners, reviewers, quality management

Deliverable

Human Judgment Review procedure

Description

Standardizes checks for sources, assumptions, client conditions, caveats, and expert judgment.

Primary users

Professionals, reviewers, managers

Deliverable

Decision Log template

Description

Records AI drafts, references, human edits, final deliverables, and approvers.

Primary users

Project teams, quality management

Deliverable

Client-contract AI-use rules

Description

Clarifies client consent, confidential information, reuse restrictions, and external AI-use permissions.

Primary users

Sales, legal, engagement owners

Deliverable

Boundary Governance rules

Description

Defines reviews when client conditions, tools, or deliverable types change.

Primary users

AI program owners, quality management, legal

Phase

1. Kickoff and scope definition

Duration

1 week

Work

Confirm target workflows, client-contract conditions, quality procedures, and assessment findings.

Output

Implementation scope

Phase

2. Decision type mapping

Duration

1-2 weeks

Work

Classify research, advice, proposals, contract review, and deliverable decisions.

Output

Professional judgment type inventory

Phase

3. Decision Boundary™ design

Duration

2-3 weeks

Work

Define AI support, human review, human final judgment, and AI-excluded decisions.

Output

Decision Boundary™ design document

Phase

4. Responsibility and review design

Duration

2 weeks

Work

Define review triggers for client submission, legal impact, financial impact, and critical proposals.

Output

Review trigger and responsibility matrix

Phase

5. Evidence design

Duration

1-2 weeks

Work

Define records for AI drafts, references, edit rationale, and final deliverables.

Output

Decision Log template

Phase

6. Governance design

Duration

1 week

Work

Define boundary review for client conditions, new tools, and deliverable types.

Output

Boundary Governance rules

Phase

7. Pilot and adoption

Duration

1-2 weeks

Work

Pilot on representative engagements and verify review burden and deliverable quality.

Output

Final deliverables and training materials

Decision type

Client-facing expert advice

AI role

Analysis and drafting support only

Review condition

Client submission, critical judgment, legal impact, or financial impact

Final owner

Engagement owner or responsible professional

Log requirement

Record AI draft, sources, human edits, and final advice

Decision type

Contract or policy review

AI role

Issue candidate generation only

Review condition

Client submission or legal impact

Final owner

Professional or legal reviewer

Log requirement

Record AI issue candidates, additional checks, and final view

Decision type

Input of client confidential information into AI

AI role

Restricted by default

Review condition

Client consent or internal approval only

Final owner

Engagement owner and legal

Log requirement

Record consent conditions, input scope, and purpose

Before

It was unclear when AI analysis became professional opinion.

After

Professional advice requires human final judgment.

Before

Contract issues not extracted by AI could be missed.

After

AI is limited to issue candidates and human additional review is required.

Before

AI use in client-facing deliverables was not retained as evidence.

After

Decision Logs record AI drafts, references, and edit rationale.

Before

AI input decisions for client confidential information were left to project teams.

After

Client consent conditions and exception approval are defined.

Before

New tools did not trigger boundary review.

After

Boundary Governance requires review when new tools or deliverable types are added.