Best when
The problem is consequential, cross-functional, and not yet ready for a responsible solution commitment.
Operating Problem Diagnostic · CG&AI
Find the smallest responsible next move for a stuck operating problem.
An Operating Problem Diagnostic is a bounded first engagement for one consequential decision, workflow, or product. CG&AI frames the outcome, organizes the evidence, exposes ownership and control gaps, and recommends whether to advise, build, operate, transfer, or stop.
Published by CG&AI · Reviewed July 12, 2026
A bounded first engagement
Diagnose before committing to a larger answer.
A complex operating problem rarely arrives in a neat category. What looks like a software request may be an ownership problem. What looks like a strategy question may depend on missing evidence. What looks ready for automation may lack an accountable approval path.
The Diagnostic creates a decision record before CG&AI or the buyer commits to a larger mandate. It separates what is known from what is assumed, identifies the people and controls the work depends on, and makes the next move explicit.
It is not a promise that technology—or CG&AI—is the answer.
Use it when
01
A consequential decision depends on evidence spread across teams, documents, tools, or advisers.
02
A workflow is important but no one owns the full path, approval logic, or exception queue.
03
A prototype, spreadsheet, or manual process cannot safely carry production responsibility.
04
The organization is debating advice, software, automation, outsourcing, or internal ownership without a common decision record.
At a glance
What the Diagnostic makes clear.
You receive
A diagnostic record, an accountable next-step plan, and a stop/go recommendation supported by the available evidence.
Possible outcomes
Advise, Build, Operate, transfer to another owner, or stop without a larger CG&AI engagement.
Choosing the intervention
Advise, Build, or Operate? Start with the next accountable output.
The modes describe different responsibilities, not a maturity ladder. Choose the smallest mode that can produce the next useful operating state with a named owner, visible evidence, and a clear completion decision.
Advise · A decision
Choose Advise when the answer is not settled.
The next accountable output is a recommendation, decision record, risk-and-condition set, or operating blueprint that leadership can approve and mobilize.
Explore Advise →Build · A production system
Choose Build when the decision is clear but the system is missing.
The next accountable output is working software, a data product, or an automated workflow with testing, controls, documentation, release, and an enduring owner.
Explore Build →Operate · Recurring performance
Choose Operate when the work must keep moving.
The next accountable output is a dependable recurring workflow with named ownership, service cadence, approval boundaries, exception handling, reporting, and a transfer path.
Explore Operate →Diagnostic · A route
Use the Diagnostic when the mode itself is still an assumption.
Frame the outcome, evidence, ownership, constraints, and stop condition before committing to advice, a build, managed operation, internal transfer, or no further work.
See the diagnostic frame →Decision gates
A problem can cross modes without becoming one open-ended engagement.
01
Name the output
Decide whether the immediate need is a decision, a production system, or recurring performance.
02
Name the owner
Identify who authorizes the work, who carries it after delivery, and who owns exceptions.
03
Set the stop condition
Define the evidence and operating state that complete the current mode before work begins.
04
Re-authorize the next mode
Build, operate, transfer, or stop only after a reviewable decision—not by default.
Diagnostic frame
Five questions before anyone prescribes a solution.
01 · Outcome
What must become true?
Define the decision or operating result, how it will be judged, and the point at which a next move is required.
02 · Evidence
What is known?
Map the available sources, contradictions, assumptions, missing facts, and evidence standards.
03 · Ownership
Who is accountable?
Identify the decision owner, operators, reviewers, technical owners, and authority boundaries.
04 · Controls
What can fail?
Expose approvals, exceptions, dependencies, reversibility requirements, and unresolved risks.
05 · Intervention
What is the smallest next move?
Choose only the work needed to reach the next defensible decision or operating state.
How it works
From an ambiguous problem to an accountable decision.
01
Fit and scope review
Confirm the problem, authority boundaries, materials, participants, timing, fee, and decision gates in writing.
02
Evidence and owner map
Organize the available sources, responsible people, dependencies, contradictions, and missing facts.
03
Constraint and control review
Make risks, approvals, exceptions, operational limits, and transfer requirements visible.
04
Stop/go recommendation
Document the smallest responsible next move, accountable owner, and conditions for proceeding.
What remains
A usable record—not a vague recommendation.
Deliverable
Diagnostic record
The problem, desired outcome, evidence base, assumptions, ownership, constraints, and unresolved questions in one traceable record.
Deliverable
Next-step plan
The smallest useful work package, its accountable owner, required inputs, control points, and completion decision.
Deliverable
Stop/go decision
A clear recommendation to advise, build, operate, transfer, or stop—without treating a larger engagement as the default.
Common questions
What buyers need to know before the first step.
What is an Operating Problem Diagnostic?
It is a bounded first engagement for a consequential decision, workflow, or product that is stuck across functions. CG&AI frames the desired outcome, organizes the available evidence, identifies ownership and control gaps, and recommends the smallest responsible next move.
When should an organization use one?
Use it when the work matters, the evidence is fragmented, ownership crosses functions, or the team cannot yet tell whether the answer is advice, a production system, a managed workflow, a controlled transfer, or no further work.
What does the buyer receive?
The buyer receives a concise diagnostic record, an accountable next-step plan, and a clear stop/go recommendation. The exact scope, timing, fee, required materials, participants, and decision gates are confirmed in writing before the engagement begins.
Does the Diagnostic automatically lead to a larger engagement?
No. A responsible outcome may be Advise, Build, Operate, a transfer to another owner, or no larger CG&AI engagement. The purpose is to make the next decision defensible, not to manufacture more work.
Does every Diagnostic involve AI?
No. The subject may be a decision, workflow, product, data problem, or operating system. When AI is relevant, the Diagnostic makes human authority, evidence standards, exception paths, and reversibility explicit rather than assuming automation is the answer.
How do we choose between Advise, Build, and Operate?
Choose by the next accountable output. Advise produces a defensible decision and mobilization path. Build produces a production system with controls and an owner. Operate produces a dependable recurring workflow with visible approvals, exceptions, and performance. If that output is not yet clear, use the Diagnostic before committing to a larger mode.
Can one problem move through more than one mode?
Yes, but each transition should be a separate decision. Advice may authorize a build; a production system may later require managed operation; an operated workflow may transfer to an internal owner. CG&AI does not treat the next mode as automatic, and the work can stop or transfer at any decision gate.
Source notes
The routing model is original; the operating principles are public.
CG&AI's Advise–Build–Operate routing model is its own engagement framework. Its underlying principles align with public guidance to understand the problem before committing to a build, assign accountable ownership, deliver iteratively, and govern AI risk throughout the lifecycle.
Reference points: GOV.UK Service Manual — how the discovery phase works; U.S. Digital Services Playbook; and NIST AI RMF Playbook. These sources inform the principles; they do not endorse CG&AI or verify its client work.
Related control guide
If AI may influence the work, govern one workflow before scaling it.
For private-market investment operations, CG&AI's five-control-point guide shows how decision ownership, evidence traceability, approval boundaries, exception handling, and transferability fit the same diagnostic frame.
Representative proof
See the kinds of operating records the work can support.
Bring one piece of work that is stuck.
Start with a fit and scope review. If the problem is not ready for a Diagnostic—or CG&AI is not the right owner—we will say so.