Gains:
- Ability to understand the concepts of change request, issue log, change control board (CCB) and quality criteria and produce an impact analysis draft with artificial intelligence support.
- Ability to use artificial intelligence to visualize the scope-time-cost-quality (iron triangle) impact of the change and draft root cause analysis
- Ability to understand that change approval and quality acceptance belong to the competent decision-maker and that artificial intelligence impact analysis must be verified.
No project goes as planned. A customer brings a new request, an unexpected error pops up, a requirement changes. The subject of this unit is to manage these inevitable changes before they turn into chaos. We will learn about three mechanisms: change management, which ensures that no work changes without approval, problem management, which records and resolves problems that arise, and quality management, which ensures that deliverables meet “good enough.” AI is a powerful analysis partner in all three: it makes visible the scope-time-cost-quality impact of a change request, investigates the root cause of problems, drafts quality criteria. But approval of change and acceptance of quality always rests with the competent decision-maker; The impact analysis of AI should not be turned into a decision without being verified.
Change management and the iron triangle
A change request is a formal request that proposes a change in scope, schedule, budget, or resource. Uncontrolled change is the main source of scope creep we've seen in previous units. The solution is to push every change through a gate: the change control board (CCB) is the authoritative group that evaluates and approves/rejects change requests.
To understand the impact of each change, the concept of the iron triangle is critical: scope, time and cost are interconnected (with quality in the middle). Changing one affects the others: if you increase the scope, either the time will increase, the cost will increase, or the quality will decrease; “more work in the same time, on the same budget” often comes at the price of quality. A good impact analysis clearly shows the impact of a change on these three (four) dimensions.
The change process is typically: request registration → impact analysis (scope/time/cost/quality/risk) → CCB decision → plan, schedule and budget update if approved → stakeholder briefing. Any unapproved changes will not be implemented.
Problem and quality management
An issue, unlike a risk, is a problem that has already occurred (risk is uncertainty in the future, problem is reality today). The issue log is a live list that tracks open issues, their priority, owner, and resolution status. Two techniques are common for finding the root cause of problems: 5 Whys — “why?” getting down to the root cause from the surface symptom by asking the question consecutively; and fishbone diagram—mapping causes into categories (human, process, material, machine, environment).
Quality management has two parts: quality assurance (QA) ensures that processes are working correctly (preventive), quality control (QC) checks whether outputs meet criteria (detector). Acceptance criteria and Definition of Done are the criteria that determine when a job is truly finished.
concept
what
example
change request
Official request that changes the plan
"Add filter to report screen"
Impact analysis
Scope/time/cost/quality impact
"+5 days, +3% budget, medium risk"
CCB
approval authority
Sponsor + PM + technical leader
problem
Realized problem
"Test environment crashed"
root cause
The real reason (5 Reasons)
"Backup configuration is incorrect"
Quality criterion
Acceptance criteria
"Error rate < 1%"
Step by step: change and quality with AI
- Clarify the request. Write the change request as "what, why, who wants it"; Ambiguous demand cannot be analyzed.
- Impact analysis draft. Ask AI for an impact outline in terms of scope, time, cost, quality and risk; verify numbers with team data.
- Generate options. Have the AI list the "approve/reject/defer/partial apply" options and the results of each.
- Submit to CCB. Take the analysis to the decision maker; Do not apply without approval.
- Root cause analysis. Have the AI generate 5 Why chains and fishbone categories for a problem; Test with real data.
- Quality criteria control. Give the deliverables to AI and have the deficiencies/non-conformities drafted according to the acceptance criteria; The final acceptance is given by the expert.
Caution: AI may make the impact of a change seem slight, such as "only 2 days", because it does not know the hidden dependencies and indirect effects. The impact analysis should not be presented to the CCB as "final" without verification with the team that will perform the work.
three mini cases
Case 1 — The real cost of change. A customer wanted a "minor screen change." PM gave the request to AI and received a draft impact analysis: the change affected three modules, +6 days and +4% budget. The team confirmed this. CCB showed the real cost to the customer; the client postponed the change to the next phase. The demand, which was thought to be "small", was managed before it turned into chaos.
Case 2 — Root cause found. In one team, the testing environment was constantly crashing. The coordinator gave the problem report to AI and asked for a 5 Why chain. The chain came down to “insufficient disks → cleanup task undefined → no process owner.” The team solved the root cause (the orphaned cleaning process), not the surface symptom (the collapse); The problem did not recur.
Case 3 — Underestimated impact. One team approved AI's "this change has minimal impact" draft without verifying it. The change broke a dependency on the critical path and the project was delayed by 9 days. Lesson: impact analysis cannot be used as a basis for decisions without team validation.
Weak prompt / Strong prompt
Weak prompt:
Consider this change request.
No size, no data and no decision framework; The AI gives a superficial and possibly overly optimistic answer.
Powerful prompt:
Your role: a change management analyst.Change request: [description]. Requested by: [role]. Justification: [why].Context: current scope, schedule (critical path attached), budget status (in ratio).Task: Impact analysis via iron triangle Produce DRAFT:- Scope impact, Time impact (will it affect the critical path?), Cost impact, Quality impact, New risks- Options: approve / reject / postpone / partial; result of eachRule: DRAFT the numeric effects and mark them with "[team verification required]".Assume you don't know the hidden dependencies; precise speech. The final decision rests with the CCB.
This prompt is powerful: it includes an iron triangle frame, option generation, draft alert, and decision maker emphasis.
Additional templates:
#5 Why engineThe question "why?" Get to the root cause by asking the question 5 times in a row: [problem]. At each step, also write how the next cause will be verified with data. Adding a made up reason.
# Fishbone producerList the possible causes of the following problem by category (Human, Process, Tool/Machine, Material, Environment, Method). Tick the 3 most likely reasons and suggest a verification method.
# Quality acceptance inspectorCheck the delivery item by item according to the following acceptance criteria; Distinguish between met, unmet and uncertain. State that the final acceptance decision rests with the expert.
Common mistakes
- Implementing the change without approval: Change without approval is scope creep itself.
- Underestimating the impact: What AI calls a "small" change can be big with hidden dependencies.
- Solving the symptom and leaving the root cause: If the 5 Whys are not done, the problem will return.
- Confusing problem with risk: Risk in the future, problem in the present; They are managed differently.
- Leaving the quality criterion subjective: "Goodness" cannot be measured; The acceptance criterion must be numerical.
- Submitting impact analysis to CCB without verification: Wrong analysis fuels wrong decision.
Tip: Saying "no" to every change request is also a management decision. A good PM knows that rejecting the change also protects the project; The PM accepts every request and manages the customer, not the project.
In summary
Change, problem and quality management keeps the project afloat in inevitable change. Changes pass through the CCB and are analyzed through the iron triangle (scope-time-cost-quality); Problems are recorded and the root cause is addressed with the 5 Whys and fishbones; Quality is assured by measurable acceptance criteria. AI accelerates impact analysis, root cause investigation, and quality auditing. However, team verification of impact numbers, change approval and quality acceptance rest with the competent human authority.
Application task
Receive a change request (actual or potential) from your project. Generate an impact analysis outline and decision options from AI through the iron triangle; verify the numbers with someone on your team. Also, take a current problem, get to the root cause with the "5 Whys engine" and direct the solution to the root cause. Summarize the impact analysis in a CCB decision format.
checklist
- [ ] I analyzed the change through the iron triangle (scope/time/cost/quality).
- [ ] I verified the impact numbers with team data, marked as draft.
- [ ] I took the change to the competent authority (CCB) for approval.
- [ ] I found the root cause of the problem with 5 Reasons/fishbone.
- [ ] I linked quality acceptance to measurable criteria.
- [ ] I did not implement any changes without approval.