Introducing our AI manifesto

Why now?

AI agents are changing how high-volume frontline hiring gets done. That pace of change demands clarity. Today we are publishing the Kiku AI Manifesto to make our principles explicit, so customers, candidates and partners know what to expect from our product and our team.

The Kiku AI Manifesto

Human-Centred

Our AI system exists to empower human hiring teams. Every system we design starts with a clear purpose to amplify human potential, foster fairness, and create meaningful outcomes. We embed ethics, inclusivity, and dignity into every layer, ensuring our technology reflects the diversity of the people it serves. Human judgment remains essential; we build agents that act, but always under thoughtful human oversight. We have hired frontline roles at a high volume, so we commit to keeping the candidate experience central to our solution.

Safety First

We hold ourselves to the highest standards of responsibility. Our systems are transparent, accountable, and compliant with evolving global regulations. We safeguard data through robust security, privacy-by-design principles, and strict governance. Every agent is tested and monitored, because we believe that the best way to earn trust is by proving our systems are safe every day.

Always Evolving

We see responsible AI as a living practice, not a static policy. Continuous feedback and real-world performance guide our improvements, keeping us aligned with our mission and values. We stay open, learning from users, peers, and partners, and we share what we discover to raise the standard for everyone. Through curiosity, reflection, and collaboration, we build AI systems that get better and more accountable over time.

What this means in practice

Here is how these principles show up in our day-to-day work.

• Human oversight as a feature, not a failsafe
• Supervisory controls, clear handoffs and audit trails so hiring teams stay in charge of decisions.
• Candidate experience as a core metric
• We measure drop-off, time to first response and candidate satisfaction, and we use that data to iterate.
• Ongoing safety testing
• Regular evaluation for bias, robustness and prompt injection, with documented corrections.
• Feedback loops everywhere
• In-product and in person feedback for customers, and a commitment to close the loop when we learn.

What's next?

Publishing a manifesto is a starting point. We will keep sharing how these principles translate into product choices and customer outcomes, and we will invite critique that helps us get better.

‍