ClarityLoop Whitepaper
The Work Is Already Speaking
Why the future of performance development will be built from the work people are already doing
Performance has been managed after the fact.
Growth is already happening in the work.
The future is to make it visible while it still matters.
Core argument
We made performance more continuous. We did not always make it easier.
Most organisations have tried to move beyond annual reviews. They added more feedback, check-ins, pulse surveys, goals and 1:1s. The intent was right: performance development should be more timely, more human and more useful.
But many processes still ask people to stop, remember what happened, write it down and explain it after the fact. Busy people struggle to do that well, even when the process is well designed.
The real opportunity is not more performance process. It is better use of the work context that already exists.
This paper argues for a simple shift: performance development should start closer to the work. Not to score people, but to help managers and employees discuss real examples while the details still matter.
Audio version
Listen to the whitepaper
A podcast-style audio version of this whitepaper, generated with AI.
Download audioIn this paper
What the paper covers
The paper follows one argument: performance development has become more frequent, but not easier to adopt. The research points to a different model, where real work context helps people grow without turning development into more admin.
01
The adoption gap
The performance adoption gap is the distance between the process an organisation designs and the process people can actually sustain while work is busy.
Memory arrives too late.
Performance systems still ask people to look backwards. By the time review season arrives, the evidence is weaker: examples are scattered, the loudest moments dominate and the most recent stories crowd out the full pattern.
That does not make people careless. It makes them human. A system built on delayed memory will always struggle with bias, incompleteness and late action.
More frequent is not always easier.
Feedback, 1:1s, pulse surveys, goals and check-ins can all help. But when they still depend on people stopping to remember, write, rate and explain, the annual form becomes a smaller form repeated more often.
Adoption breaks in predictable places.
- Employees rush feedback because the request arrives outside the flow of work.
- Managers prepare from memory because examples are scattered across tools and moments.
- HR spends too much energy chasing completion instead of improving the quality of conversations.
- Leaders see signals after the moment for useful action has already passed.
The adoption gap is not a side issue. It is one of the main reasons good performance ideas fail.
02
Why now
Organisations need earlier development signals at the exact moment managers, employees and HR teams have less patience for process that does not help them act.
Engagement and skill pressure are converging.
Gallup's State of the Global Workplace 2026 reports that global employee engagement fell to 20% in 2025, its lowest level since 2020, and estimates the productivity cost of low engagement at about $10 trillion. Manager engagement fell from 31% in 2022 to 22% in 2025.
The World Economic Forum's Future of Jobs Report 2025 says employers expect 39% of workers' core skills to change by 2030. It also reports that 63% of employers identify skill gaps as a major barrier to business transformation.
20%
global employee engagement in 2025, Gallup's lowest level since 2020
22%
global manager engagement in 2025, down from 31% in 2022 according to Gallup
39%
of workers' core skills expected by employers to change by 2030, according to WEF
63%
of employers identify skill gaps as a major barrier to business transformation, according to WEF
The old answer is not enough.
These pressures make timely development more important, not less. People need feedback while it can still change the work. Managers need earlier signals before small issues become larger patterns. HR needs processes people can actually adopt.
Adding more forms, reminders and survey cycles is not enough if those processes still depend on people reconstructing work from memory.
03
The thesis
Growth is not separate from work. It happens inside the work.
The work already contains development signals.
People grow when they ask better questions, respond to feedback, support a colleague, resolve a blocker, explain something clearly, recover from a mistake, improve a handover, coach someone else or take more ownership.
The signals are often ordinary moments:
- a thoughtful review comment
- a repeated blocker
- a difficult handover
- a moment of support
- a missed follow-up
- a pattern of unclear ownership
- a sign that someone is stepping up
Most performance systems are not designed to notice those moments. They wait until later and ask people to summarise what they remember.
Context-first development starts closer to reality.
Context-first performance development means using real work context to support better human development. It does not mean every message, meeting or task becomes a performance record. It does not mean software decides who is good or bad. It does not mean managers are replaced.
It means the system helps surface recent, relevant patterns so employees and managers can have more specific development conversations.
This is bigger than reviews.
Reviews are one use case, but not the main idea. A context-first model can support feedback, 1:1s, recognition, manager preparation, growth planning, career development, engagement listening and team health conversations.
It also makes engagement listening more grounded. Surveys still matter because employees need a direct voice. But surveys should not be the only way an organisation understands blockers, friction, unclear ownership or missing recognition.
The practical result should be less friction, clearer examples and earlier action.
04
The evidence
CollabSense gives early evidence for the central thesis: useful development signals can come from the work itself.
We tested whether work can reveal growth signals.
CollabSense is an open research study created by ClarityLoop with Edinburgh Napier University. It looked at public collaboration data from real open-source software projects to test a simple question:
Can the way people work together reveal useful signals about strengths, collaboration and growth opportunities?
This matters because performance development usually depends on people reporting what happened after the fact. CollabSense tested whether the work itself could provide a better starting point.
3,300+
public GitHub interactions scored in the CollabSense research study
6.63/10
mean sentiment score across the analysed interactions, concentrated in constructive ranges
85%
sampled growth opportunities judged at least partially actionable
90%
sampled growth opportunities judged at least partially evidence-supported
The answer was encouraging.
Across more than 3,300 real collaboration interactions, CollabSense was able to surface strengths and potential growth opportunities from the way people worked together.
The strongest signals appeared when there was enough repeated context. A single comment rarely tells you much. But patterns across collaboration can show how someone communicates, supports others, follows through, responds to feedback and grows over time.
Work carries signal
The way people collaborate can reveal useful patterns about strengths, support, communication and growth.
Patterns matter
The clearest signals came from repeated collaboration, not isolated moments.
Context builds trust
When there was not enough evidence, the system produced fewer useful signals. That restraint matters.
Evidence improves growth
The best use of AI is to bring forward examples people can review, discuss and learn from.
The research also showed an important boundary.
When there was not enough context, the system produced fewer useful growth signals. That is important. A trustworthy development system should not force a conclusion when the evidence is too thin.
The study also showed that asking AI to produce more output can create more noise. The better path is not more feedback for the sake of it. It is better signals from better context.
Human review supported the usefulness of the signals.
Human reviewers assessed a sample of the growth opportunities. In that sample, 85% were judged at least partially actionable and 90% were judged at least partially supported by evidence.
That gives the research practical weight. The signals were not just abstract AI outputs. Many were understandable, reviewable and linked back to real examples of work.
CollabSense shows that the work itself can become a starting point for continuous development.
06
Guardrails
A context-first model needs clear boundaries so evidence supports development instead of becoming surveillance.
- Do not turn every work signal into a performance record.
- Do not treat communication volume as performance.
- Do not infer disengagement from quietness.
- Do not let AI score people or replace manager judgement.
- Do not hide the evidence behind a summary.
- Do not force a signal when there is not enough context.
The goal is not to build a record against someone. The goal is to reduce the burden of memory and make development conversations more useful.
For this to work, the evidence must be inspectable. The purpose must be developmental. The person must not be reduced to a score. The system must be able to say when there is not enough evidence. Humans must remain responsible for judgement.
07
Design principles
The next generation of performance development should make adoption easier while keeping judgement human.
Start from the work
Connect development to real examples, not only retrospective summaries.
Reduce memory burden
Do not make growth conversations depend on what someone remembers months later.
Show evidence
A signal should be inspectable. People need to see why it appeared.
Support growth
The purpose is development, support and earlier action, not hidden monitoring.
Keep humans accountable
AI can structure context. People remain responsible for judgement.
Make adoption easier
The system should reduce friction, not add another layer of admin.
The future
From performance cycles to development in the flow of work
The industry was right to move away from annual rituals. But continuous performance will not work if it asks people to do more performance admin more often.
The deeper shift is from retrospective performance management to context-first performance development: from memory to evidence, from delayed reflection to timely signal, and from manager guesswork to clearer preparation.
The work is already speaking. The next challenge is to listen responsibly.
Not to judge people automatically. Not to remove human judgement. But to help people grow from the real context of their work.
Closing thought
Start from the work.
Performance development becomes easier to use when it starts with examples people recognise from their day-to-day work.
The point is not to score people, replace managers or turn every interaction into a record. It is to make the next conversation more specific, fair and useful.