01 / Our culture / Founding intentions
Psychological safety.
Exceptional standards.
We are intentionally building a company where trust, meaningful ownership and curiosity help people do exceptional work. High standards and respect belong together.
Choose a lens to explore the philosophy, or simply continue reading.
People need enough trust to surface a concern before it becomes a failure, and enough ownership to act on what they learn. Different perspectives make the work stronger when people can disagree without having to defend their worth.
We intend to make asking for help, giving credit and challenging an assumption ordinary parts of the work. Psychological safety protects honest participation; it does not remove responsibility for outcomes.
Showing The People
Explore trust, product thinking and leadership through a short optional path.
People need enough trust to surface a concern before it becomes a failure, and enough ownership to act on what they learn. Different perspectives make the work stronger when people can disagree without having to defend their worth.
We intend to make asking for help, giving credit and challenging an assumption ordinary parts of the work. Psychological safety protects honest participation; it does not remove responsibility for outcomes.
Ambition needs discipline. We want clear problems, thoughtful decisions and working products that earn people's trust. An experiment is useful when it changes what we understand, even when it does not produce the result we hoped for.
We distinguish exploration from a release commitment, name the owner of a decision and agree on what evidence would change our minds. Quality includes reliability, security and the experience people actually have.
A great product should make something meaningful easier for the people using it. A great team should make its members more capable, not more dependent on a single person's answers.
We intend to recognize the people who make others effective: the colleague who exposes a risk, the engineer who improves reliability and the leader who creates room for someone else to grow.
02 / The values network
Trust, curiosity, accountability, growth, psychological safety and excellence form a system. A value becomes useful when it changes behavior, especially when two commitments pull in different directions.
Select one value to explore its behaviors. Select a second to see how they work together. Selecting a third replaces the older selection.
Selected: Trust
Confidence grows when words, decisions and follow-through agree. Trust makes honest information available early.
In practiceA delivery date slips. The owner explains what changed, the remaining risk and a revised plan before a partner has to ask.
Connects to Curiosity, Accountability, Psychological Safety.
Select another value to explore a relationship or a tension.
Confidence grows when words, decisions and follow-through agree. Trust makes honest information available early.
In practiceA delivery date slips. The owner explains what changed, the remaining risk and a revised plan before a partner has to ask.
Connects to Curiosity, Accountability, Psychological Safety.
Ask what we might be missing. Treat a strong opinion as a starting point for investigation, not a substitute for evidence.
In practiceFans ask for a feature. Instead of immediately building it, the team investigates the problem that request is trying to solve.
Connects to Accountability, Growth, Excellence.
Own the decision, its consequences and the next useful action. Responsibility is clearer when authority and expectations are explicit.
In practiceAn experiment misses its goal. Its owner reports the findings, explains what was learned and proposes whether to iterate or stop.
Connects to Trust, Curiosity, Excellence.
Build capability through thoughtful stretch, feedback and reflection. One setback should inform someone's development rather than define their future.
In practiceA new lead takes on a difficult decision with a clear mandate, a coach and an agreed point to review what happened.
Connects to Curiosity, Psychological Safety, Excellence.
People can ask, disagree and report mistakes without humiliation or retaliation. Safety enables candor; it does not excuse avoidable harm.
In practiceA junior teammate identifies a release risk. The team investigates the evidence rather than dismissing the concern because of seniority.
Connects to Trust, Accountability, Growth.
Set a meaningful quality bar and improve the way we meet it. Excellence is sustained care in the work, not a demand for performative perfection.
In practiceThe team reduces launch scope to protect a reliable experience, then explains the tradeoff clearly to stakeholders.
Connects to Curiosity, Accountability, Growth.
03 / Trust and standards lab
Psychological safety is not permission to ignore the quality bar. A meaningful standard is not permission to make people afraid. Explore how the two commitments affect a team's ability to learn.
Adjust either slider. The point and explanation update together; the numbers represent positions in this illustration, not scores.
People can challenge assumptions, share concerns and learn from mistakes while remaining accountable for meaningful outcomes. Respect and a clear quality bar make difficult conversations useful.
A useful next move: Keep expectations explicit. Invite risks early, review the work fairly and make learning part of delivery.
Conceptual illustration of team dynamics, not a scientific model, validated psychological assessment or judgment of a visitor.
Unclear expectations and low confidence in the team make it hard to invest in shared outcomes. People can retreat into doing the minimum or waiting for direction.
Start with a small credible commitment, a named owner and a respectful review. Rebuild trust through follow-through.
People may feel supported, but unclear expectations can leave important work unfinished. Avoiding a difficult conversation can quietly shift the burden to other teammates.
Preserve care while clarifying outcomes, ownership and the feedback needed to improve.
The bar is high, but people may hide mistakes or stay silent to protect themselves. Problems arrive late, and apparent agreement can conceal risk.
Maintain the quality bar. Change how concerns are received, how decisions are explained and how mistakes are reviewed.
People can challenge assumptions, share concerns and learn from mistakes while remaining accountable for meaningful outcomes. Respect and a clear quality bar make difficult conversations useful.
Keep expectations explicit. Invite risks early, review the work fairly and make learning part of delivery.
04 / The decision room
Incomplete information, limited capacity and competing commitments are normal parts of building. Good judgment means making the tradeoff visible, choosing an owner and being willing to revise a decision when the evidence changes.
Choose a situation, make an initial decision, then respond to new information. Explore alternate paths without a score.
Situation 1 / 5
A teammate questions a product direction after the team has committed to it. Work is underway, and a partner is expecting a demonstration. The concern is specific: the current flow may exclude a group of fans we intend to serve.
The tensionDecision stability matters. So does new evidence. Reopening every preference creates churn; closing every discussion hides risk.
Illustrative workplace situation.Choose an initial decision.
A teammate questions a product direction after the team has committed to it. Work is underway, and a partner is expecting a demonstration. The concern is specific: the current flow may exclude a group of fans we intend to serve.
Decision stability matters. So does new evidence. Reopening every preference creates churn; closing every discussion hides risk.
Invite challenge, assess its substance and make the decision legible. The owner explains what changed, what stayed fixed and why. Respectful dissent should not become a test of loyalty.
A launch is approaching. The team finds an intermittent defect that could expose private conversation content to the wrong viewer. Stakeholders have planned an announcement, and the team is exhausted after preparing for release.
A date is a commitment, but so is protecting people. Schedule pressure cannot settle a privacy risk that has not been understood.
The release owner is accountable for protecting users, not just hitting a date. Pause unsafe scope, involve engineering and security judgment, verify the repair and communicate the tradeoff without blaming the person who raised it.
A teammate leads a bounded experiment based on an agreed hypothesis. The prototype does not help participants as expected. The work used meaningful time, and someone asks whether the owner was the wrong person for the task.
Learning has a cost. Accountability means distinguishing a thoughtful test from avoidable neglect and using the result to make a better decision.
A thoughtful experiment can fail without its owner being a failure. Review decision quality and execution honestly, make the next investment decision from evidence and share the learning so others do not have to repeat it.
A highly capable teammate delivers important work but repeatedly dismisses colleagues in reviews. Several people have stopped offering ideas. The teammate says bluntness is necessary to maintain standards.
Individual output matters, but so does the capacity of the whole team. Candor about work is different from contempt for people.
Respect is part of performance. Give specific feedback, listen for context, support change and set a clear review point. Protect teammates from repeated harm, and apply standards consistently regardless of technical capability.
A compelling discovery feature could create a strong demo. Meanwhile, an existing experience fails unpredictably. The same engineers are needed for both, and a partner is interested in seeing something new.
Visible novelty can earn attention. Reliability earns continued trust. Capacity is finite, so the team needs an explicit tradeoff rather than two incompatible commitments.
Use the seriousness and reach of the reliability issue to guide priority. Learn about new value with the smallest credible experiment. Renegotiate scope transparently instead of letting hidden overload make the decision.
05 / The product iteration lab
Shipping another feature is not the same as solving a problem. We use evidence and real user behavior to guide what we build, including when the useful next decision is to stop.
Follow five decisions in a hypothetical product discussion. Use Back to change a choice; Restart clears this exercise.
Imagine fans participating during games but returning less often between them. The team proposes a new engagement feature. Before committing, it needs to understand the problem and learn whether a smaller experiment helps.
All findings below are invented for this exercise. They are not product claims, actual research results or a promise of public availability.
Step 1 of 5: Hypothesis. What might explain the gap?
Selected: Fans need a reason to continue the conversation. Continue to see how this changes the next decision.
Investigate why a fan would return, observe the journey and test one small response to a clear hypothesis. A shared recap might invite replies; a discovery path might help find a relevant story. A reminder alone cannot establish whether the destination offers value.
In these fictional prototype sessions, fans used a shared recap when it carried a friend's take and offered a clear way to reply. A generic recap did not create the same reason to return.
Decide whether to iterate, expand cautiously or stop. Preserve the findings and define the next learning question. Fictional prototype signals cannot demonstrate actual adoption or retention.
06 / The feedback studio
Feedback should help someone see an issue clearly enough to do something about it. Care includes telling the truth; candor includes making it useful.
Select an example, then compare possible formulations. These are illustrations, not scripts to repeat regardless of context.
The unhelpful version
“Your presentation wasn't good enough.”
A product review needed a clear decision, but the recommendation and evidence were difficult to find.
Useful feedback makes the issue specific enough to act on, asks for context and arrives while improvement is still possible. Choose a formulation to examine its effect.
Have the conversation in a setting suited to the issue. Respect does not require avoiding the difficult part, and candor does not require humiliating someone.
Describe an observable behavior and its impact. Distinguish a pattern from an absolute like “always.” Ask for context, agree on a next step and return to the conversation while the work can still improve. Receive feedback by clarifying the observation before defending an intention.
07 / Human judgment × agentic development
We use AI agents to accelerate development, test ideas and continuously improve products. Experienced people own the vision, decisions, quality and consequences. Follow an idea from problem definition to learning.
Compare conventional and AI-assisted development, then select a stage to examine ownership. The comparison changes responsibilities and tools, not invented speed or productivity numbers.
Stage 1 / 8
Tools change.
Ownership does not.
AI-assisted approach: Define the problem
People own: Choose who we serve, the problem worth solving and the boundaries of acceptable risk.
Agents can assist: Help organize existing evidence and identify unanswered questions.
A fluent summary is not validation. Confirm sources and speak with the people affected.
People own: Own priorities, tradeoffs, scope and the decision to invest.
Agents can assist: Generate alternatives and surface assumptions to challenge.
Tools can propose options; they cannot own the consequences of a product decision.
People own: Define research questions, verify sources and interpret limitations.
Agents can assist: Assist with research synthesis, prototypes and test ideas.
Treat agent output as material to verify. Respect rights, consent and data boundaries.
People own: Own architecture, interfaces and engineering decisions.
Agents can assist: Implement bounded tasks, propose changes and explain their assumptions.
Constrain access and scope. Review dependencies, sensitive data handling and generated changes.
People own: Define meaningful acceptance criteria and investigate failures.
Agents can assist: Help write and run checks; automation executes repeatable tests.
Passing tests only establish what those tests cover. Avoid checks that merely echo the implementation.
People own: Review behavior, quality, privacy and security; approve or reject the change.
Agents can assist: Assist with finding risks and explaining code paths.
Human ownership remains explicit. An agent reviewing another agent's output is not a transfer of accountability.
People own: Observe actual use and decide whether the experience solves the intended problem.
Agents can assist: Help organize feedback and compare findings with the original assumptions.
Do not replace real participants with simulated approval or confuse a prototype signal with proven adoption.
People own: Decide what to improve, expand or stop and communicate the reasons.
Agents can assist: Help investigate issues, implement refinements and rerun verification.
A shorter path to an implementation does not remove the need for judgment, verification or learning.
08 / The leadership explorer
Leading well makes other people more effective. It creates clarity, develops judgment and gives people real authority, while remaining accountable for the conditions in which they work.
Choose a commitment to compare an ineffective behavior with a practical alternative. The reflection is yours; nothing is recorded.
Instead of
Provide instructions without explaining the problem, then change the goal midstream.
We believe in
Share the goal, constraints and why the work matters; agree on how progress and risks will be reviewed.
In practiceBefore a prototype starts, the lead defines the question it must answer and the quality boundaries. The team chooses how to explore it.
A question to carry with youCould someone explain both the outcome and what they are empowered to decide?
Make the outcome, constraints and decision owner clear enough that people can act without asking permission for every move.
Share the goal, constraints and why the work matters; agree on how progress and risks will be reviewed.
Before a prototype starts, the lead defines the question it must answer and the quality boundaries. The team chooses how to explore it.
Help people strengthen their judgment. Support should make them more capable, not dependent on your answer.
Ask how someone sees the problem, offer specific feedback and agree on a supported next step.
A new owner presents alternatives and risks. The lead tests the reasoning, helps identify gaps and leaves the decision authority intact.
Say what needs to be said while respecting the person receiving it. Clarity and kindness work together when feedback is specific.
Describe the behavior, its impact and a useful next action; listen for context.
After a missed handoff, discuss the agreed date and the blocked team, then work out an early risk update and a realistic recovery plan.
Give people real authority with context, support and responsibility for the outcome.
Agree on the outcome and boundaries, then let the owner decide within them and explain the result.
A release owner can reduce scope when acceptance criteria are not met, while keeping partners informed about the tradeoff.
Leadership creates conditions for others to succeed. Make meaningful contributions visible, including the quiet work that protects quality.
Give credit generously and take responsibility personally for decisions you own.
A launch review names the people who found risks, improved reliability and supported others, not just the person presenting the demo.
09 / The team agreement
Trust is reciprocal. These founding commitments describe the team we intend to build and the behavior we should be willing to discuss honestly when we fall short.
Switch perspectives to see reciprocal commitments. These describe founding intentions, not proven company-wide policies or contractual terms.
Explain the priorities, constraints and reasons behind decisions. Distinguish what is known from what remains uncertain.
Give timely, specific feedback; listen when someone challenges a decision or reports a concern.
Pair responsibility with decision authority, support and a clear path for escalation.
Offer supported stretch, coaching and honest reflection. Do not make one thoughtful mistake a permanent label.
Plan for meaningful quality and recovery time. Do not treat chronic overload as proof of commitment.
Apply expectations consistently, give credit and take responsibility for leadership decisions.
Showing What teammates can expect from leadership
Context and transparency: Explain the priorities, constraints and reasons behind decisions. Distinguish what is known from what remains uncertain.
Respectful candor: Give timely, specific feedback; listen when someone challenges a decision or reports a concern.
Real ownership: Pair responsibility with decision authority, support and a clear path for escalation.
Room to grow: Offer supported stretch, coaching and honest reflection. Do not make one thoughtful mistake a permanent label.
Sustainable standards: Plan for meaningful quality and recovery time. Do not treat chronic overload as proof of commitment.
Fair accountability: Apply expectations consistently, give credit and take responsibility for leadership decisions.
Follow-through: Own commitments, surface risks early and renegotiate scope honestly when the facts change.
Productive disagreement: Challenge the reasoning with evidence and respect. Once a decision is made, act on it or raise a material new concern clearly.
Care for the whole team: Share context and credit. Avoid behavior that makes others less able to contribute.
Learning and feedback: Ask for help, offer specific feedback and use setbacks to improve the work.
Quality and responsible AI: Verify outputs, protect sensitive information and remain accountable for changes made with tools or agents.
Honest capacity: Make limits visible rather than relying on hidden overtime. Protect the team's ability to perform well over time.
10 / Growing intentionally
A founder-led company needs its culture to become more explicit as it grows. These future stages are aspirations. They describe challenges we want to anticipate, not achievements we claim today.
Select a stage to explore a cultural risk and an intended response. This is a development model, not a record of milestones achieved.
Founding intent
Selected stage: Founder-led beginnings. Founding intent.
Too much context and authority can stay in the founders' heads. Speed becomes fragile when every question needs the same person.
Name owners, write down the reasoning and make uncertainty visible.
Use small decision records, explicit quality criteria and regular conversations about what is working. These are intended practices, not a claim that an established operating system is already in place.
Early relationships can turn into unwritten rules that new teammates cannot see. Hiring for agreement can reduce the range of useful perspectives.
Make expectations and decision boundaries explicit; welcome different ways of seeing the problem.
Introduce reciprocal commitments, supported onboarding and feedback that helps both newcomers and founders improve.
Delegation can leave responsibility without authority. Local optimizations can make coordination harder across teams.
Develop leaders who explain context, coach judgment and coordinate dependencies without taking over every decision.
Clarify ownership across interfaces, review decision quality and create opportunities for emerging leaders to practice with support.
Process can become ceremony, and cultural slogans can outlive the behaviors they were meant to describe.
Keep asking whether the operating practices support trust, quality and useful autonomy as the organization changes.
Review practices against real experience, make concerns safe to surface and revise the system when evidence shows it is no longer helping.
11 / A founder perspective
The principles here reflect Katwalk Labs' founding direction. This is an editorial perspective, not an invented first-person quotation or recorded testimonial.
Chi Owunwanne
Founder & CEO
Chi brings two decades of Microsoft product and platform experience across Xbox, Windows, Surface and Azure Virtual Desktop. Leading complex technology work requires more than an ambitious idea: people need shared context, clear ownership and enough trust to reveal a problem while it can still be addressed.
That perspective informs the company Katwalk Labs intends to build. Psychological safety and high standards are complementary commitments. People should be able to challenge an assumption, ask for help and acknowledge a mistake; the team should still expect careful engineering, honest communication and responsibility for what ships.
The aim is not a team that agrees about everything. It is a team that can turn different perspectives into better judgment, make difficult tradeoffs without losing respect and learn from the result. Leadership helps create those conditions, then makes room for others to exercise their own judgment.
Agentic development makes this responsibility more explicit. Tools can help explore, implement and verify. They do not replace the people accountable for the problem being solved, the risks being accepted or the experience someone receives.
Meet Chi and Enyinne →12 / Our commitment
Give credit generously. Take responsibility personally.
Keep learning. Help others grow.
This page expresses founding principles and intended commitments. It does not represent a visitor assessment, employment policy or evidence of company-wide outcomes. Your selections are held only in page memory and are cleared when this page reloads.