Current Pattern → Value Potential #1
Annual Performance Review → Continuous Performance Development
A Pattern Value Engineering (PVE) Analysis
Most organizations do not deliberately design work from first principles.
They inherit patterns.
A meeting happens every Monday because it has always happened every Monday.
A report is produced monthly because monthly reporting became the accepted pattern.
Performance is reviewed annually because organizations developed systems around annual appraisal cycles.
The arrival of AI gives us an opportunity to ask a different question:
Should we use AI to make the current pattern faster—or does AI make a better pattern possible?
That is the purpose of this new Current Pattern → Value Potential series.
Each post takes a familiar pattern of work and examines it through Pattern Value Engineering (PVE):
Current Pattern → Why This Pattern? → Value Leak → PVE Move → AI Lever → Pattern Shift → Value Potential → Human Reserved → Proof → New Pattern
And importantly:
Value Potential is not Verified Value.
Until an intervention is implemented and measured, we should call it potential.
Our first pattern:
THE ANNUAL PERFORMANCE REVIEW
1. CURRENT PATTERN
A familiar organizational pattern looks like this:
Set Goals → Work for Months → Gather Information → Manager Reviews → Annual Conversation → Rating → Development Plan → Repeat
There are many variations.
Some organizations use quarterly reviews.
Some combine annual reviews with informal conversations.
Some have moved toward continuous feedback.
Structured reviews themselves are not necessarily the problem. CIPD's 2026 guidance notes that, although annual reviews have increasingly been questioned in favour of more regular conversations, structured performance reviews can still form an important part of a broader performance-management cycle.
The Pattern AI question therefore isn't:
“Are annual reviews bad?”
It is:
Why did this pattern emerge, where does it create value, where does it leak value, and what parts of it could now be redesigned?
2. WHY THIS PATTERN?
The annual review solves genuine organizational problems.
Organizations need mechanisms to:
clarify expectations,
review performance,
document progress,
identify development needs,
support accountability,
and inform consequential decisions.
But continuously observing, remembering, synthesizing and discussing every employee's work is cognitively and administratively expensive.
So organizations compress performance information.
Months of work become:
documents,
ratings,
manager recollections,
employee self-assessments,
and periodic conversations.
In Pattern Value Engineering terms, the annual review can therefore be viewed partly as a:
COMPRESSION PATTERN.
A large amount of activity is periodically compressed into something manageable.
That made practical sense.
But AI changes the economics of:
capturing,
retrieving,
summarizing,
comparing,
and synthesizing information.
And that means the old compression pattern deserves another look.
3. VALUE LEAK
Where might value be leaking from the existing pattern?
The first issue is:
DELAY.
Feedback about something that happened months ago has less opportunity to influence what happens tomorrow.
Gallup has reported that only 14% of employees strongly agree their performance reviews inspire them to improve.
It has also found that employees receiving weekly rather than annual feedback are substantially more likely to report receiving meaningful feedback and being motivated to do outstanding work.
More recent Gallup research continues to emphasize frequent, meaningful performance conversations.
The second issue is:
MEMORY.
A manager must reconstruct months of performance.
Recent, visible or memorable events may disproportionately influence judgment.
The third is:
ADMINISTRATIVE EFFORT.
Managers and employees spend time collecting information that already exists across:
projects,
goals,
work systems,
documents,
feedback,
customer interactions,
and other records.
The fourth is:
LOST LEARNING TIME.
If an important performance pattern appears in February but is meaningfully discussed in December, months of possible improvement have disappeared.
The fifth is:
EVENT-BASED DEVELOPMENT.
Development becomes something discussed during the appraisal rather than something embedded continuously in work.
These are potential Value Leaks.
But this does not mean replacing the manager with AI.
Quite the opposite.
It creates an opportunity to use AI to reduce information friction so managers can spend more time doing what machines cannot responsibly replace.
4. THE PVE MOVE
Which of the six Pattern Value Moves applies?
There are actually several.
RECOGNIZE
Identify recurring patterns in:
goals,
progress,
strengths,
barriers,
feedback,
learning needs,
and outcomes.
COMBINE
Bring together relevant evidence that currently sits across different work contexts.
IMPROVE
Reduce the administrative burden of preparing for performance conversations.
But the most interesting move is:
REIMAGINE.
Instead of asking:
How can AI make the annual appraisal faster?
ask:
What would performance development look like if relevant evidence could be synthesized continuously?
That changes the problem.
5. AI LEVER
AI could potentially assist with activities such as:
summarizing agreed work evidence,
tracking progress against explicit goals,
organizing employee-provided accomplishments,
identifying recurring themes in feedback,
preparing questions for coaching conversations,
surfacing unresolved commitments,
drafting development summaries,
and helping managers prepare for conversations.
The important word is:
ASSIST.
AI should not silently become the judge of the employee.
There is an enormous difference between:
AI helping humans understand evidence
and
AI deciding the worth of a human being.
PVE must preserve that distinction.
6. PATTERN SHIFT
The current pattern might look like:
Annual Goals
↓
Months of Work
↓
Memory + Documents
↓
Annual Synthesis
↓
Manager Judgment
↓
Feedback
↓
Action
A redesigned pattern could become:
Work
↓
Relevant Evidence
↓
AI-Assisted Synthesis
↓
Regular Human Conversation
↓
Manager + Employee Judgment
↓
Timely Action
↓
Learning
↓
Updated Goals
Then repeat.
This is not:
Continuous Evaluation.
It should be:
Continuous Performance Development.
That distinction matters enormously.
7. VALUE POTENTIAL
What value might the new pattern create?
TIME VALUE
Less manual preparation before formal reviews.
LEARNING VALUE
Feedback can arrive closer to the event.
QUALITY VALUE
Performance conversations can potentially use broader evidence rather than recollection alone.
DEVELOPMENT VALUE
Patterns of strength and difficulty may become visible earlier.
MANAGERIAL VALUE
Managers can spend less time reconstructing history and more time:
coaching,
understanding,
supporting,
and deciding.
EMPLOYEE VALUE
Employees may receive greater clarity about:
what is working,
what needs attention,
and what to do next.
ORGANIZATIONAL VALUE
Performance management can potentially move from:
administrative event
toward:
continuous capability development.
But these are still:
VALUE POTENTIALS.
They have not yet been proved.
8. HUMAN RESERVED
This is one of the most important PVE questions:
What should remain human?
AI may help organize evidence.
But humans should remain responsible for consequential judgments involving:
context,
fairness,
intent,
circumstances,
potential,
empathy,
career aspirations,
conflict,
exceptions,
development,
and accountability.
AI may say:
“This pattern appears repeatedly.”
A manager must ask:
“Why?”
AI may summarize what happened.
A human conversation must understand:
what it meant.
AI may help prepare feedback.
The manager remains responsible for:
the relationship.
9. THE RISK: FROM DEVELOPMENT TO SURVEILLANCE
There is another possible future.
Organizations could use AI not to support people but to monitor them continuously.
That would create a very different pattern:
Employee → Continuous Monitoring → Algorithmic Evaluation → Automated Judgment
The OECD has warned that algorithmic management and pervasive monitoring can increase work intensity, reduce autonomy and create stress.
That means:
More data does not automatically create more value.
PVE must ask about:
value created,
value destroyed,
and value shifted between stakeholders.
An AI system that saves managerial time but destroys employee trust may have negative overall value.
Therefore the new pattern must be deliberately designed around:
transparency,
appropriate data boundaries,
human oversight,
employee agency,
and legitimate organizational purpose.
10. PROVE
How would we know whether the new pattern actually creates value?
Start with a baseline.
Measure the current system.
Possible measures include:
Manager preparation hours per employee
Employee preparation hours
Frequency of meaningful feedback
Time between an event and useful feedback
Employee perception of feedback usefulness
Goal clarity
Development actions completed
Manager perception of conversation quality
Employee perception of fairness
Performance improvement where measurable
Retention or engagement where relevant
Trust and perceived surveillance
Then introduce the redesigned pattern.
Measure again.
Only then can we move from:
VALUE POTENTIAL
to:
VALUE REALIZED.
And only with appropriate evidence should we call it:
VERIFIED VALUE.
11. THE PVE SCORECARD
VALUE POTENTIAL
High
Performance management affects development, managerial time, employee experience and organizational capability.
AI–PATTERN FIT
Potentially High for synthesis and assistance; much lower for autonomous judgment
AI is well suited to organizing, retrieving and summarizing relevant information, but consequential employee evaluation requires strong human oversight.
HUMAN JUDGMENT REQUIREMENT
Very High
Performance management involves context, fairness, relationships, development and consequential decisions.
EVIDENCE STRENGTH
Moderate for the benefits of frequent meaningful feedback; emerging and context-dependent for AI-enabled redesign
Research supports the importance of effective, timely feedback, but that does not automatically establish the superiority of every AI-enabled continuous performance-management system.
That distinction matters.
12. CURRENT PATTERN → NEW PATTERN
CURRENT
Periodic Evidence
→ Retrospective Review
→ Rating
→ Feedback
→ Development
POSSIBLE NEW PATTERN
Ongoing Work
→ Relevant Evidence
→ AI-Assisted Synthesis
→ Human Conversation
→ Timely Intervention
→ Learning
→ Adaptation
The change is not merely:
Annual Review → AI Review.
The more interesting shift is:
Performance Evaluation → Performance Development.
13. The Larger Pattern
This first example reveals something important about Pattern Value Engineering.
The annual performance review is not fundamentally an AI problem.
It is a:
PATTERN PROBLEM.
AI simply changes what patterns are economically and cognitively possible.
The old question is:
How can AI improve performance reviews?
The PVE question is:
What outcome was the performance review designed to accomplish, and what is now the best pattern for accomplishing it?
That is a very different question.
And it leads to a very different AI strategy.
The PVE Lesson
Whenever you encounter an established process, don't immediately ask:
“How can AI automate this?”
Ask:
What is the current pattern?
Why does this pattern exist?
Where does it leak value?
Which old constraint created it?
Has AI changed that constraint?
What new pattern becomes possible?
What value might the new pattern create?
What must remain human?
How would we prove the value?
That is:
CURRENT PATTERN → VALUE POTENTIAL.
And ultimately:
Current Pattern
→ Value Leak
→ Pattern Shift
→ Value Potential
→ Intervention
→ Value Realized
→ Evidence
→ Verified Value
Evidence Base
CIPD — Performance Reviews, 2026
CIPD notes that the value of annual reviews has been questioned in favour of more regular conversations, while also emphasizing that structured performance reviews remain an important part of a holistic performance-management cycle.
https://www.cipd.org/en/knowledge/factsheets/appraisals-factsheet/
CIPD — Performance Feedback: An Evidence Review
CIPD's evidence review finds substantial research supporting feedback as a mechanism for improving performance, while also warning that poorly designed feedback can be detrimental.
https://www.cipd.org/en/knowledge/evidence-reviews/performance-feedback/
Gallup — The Truth About Performance Reviews
Gallup reports that only 14% of employees strongly agree that their performance reviews inspire them to improve and presents evidence supporting more frequent, meaningful manager-employee conversations.
https://www.gallup.com/workplace/249332/harm-good-truth-performance-reviews.aspx
Gallup — How Effective Feedback Fuels Performance
Gallup reports substantially stronger employee responses where feedback is frequent rather than annual and emphasizes ongoing coaching conversations.
https://www.gallup.com/workplace/357764/fast-feedback-fuels-performance.aspx
OECD — Artificial Intelligence, Job Quality and Inclusiveness
The OECD discusses risks associated with algorithmic management, including increased work intensity, pervasive monitoring and reduced worker autonomy.
Evidence Note
This post does not claim that replacing annual performance reviews with AI will automatically improve performance.
The evidence supports the importance of timely, meaningful feedback and identifies limitations in traditional performance-management practices.
The proposed AI-enabled pattern is a PVE hypothesis about Value Potential.
Its actual value must be established through implementation, comparison, employee safeguards and evidence.
That distinction is fundamental to Pattern Value Engineering.
Series Framework
CURRENT PATTERN — How is it accomplished today?
WHY THIS PATTERN? — What constraint created it?
VALUE LEAK — Where is value lost or unrealized?
PVE MOVE — Recognize / Replicate / Improve / Predict / Combine / Reimagine.
AI LEVER — What does AI make possible?
PATTERN SHIFT — What changes?
VALUE POTENTIAL — What could become better?
HUMAN RESERVED — What remains under human judgment?
PROVE — What evidence would establish value?
NEW PATTERN — What does the redesigned accomplishment look like?
Current Pattern → Value Potential #1
Pattern Value Engineering (PVE) v0.1
AI YING — Accomplishing with AI
