Humaneer

Humaneer already had a product and users. What they needed was practical AI capability layered into it: not a separate tool, and not a demo.

HR and compliance is a domain where getting it approximately right is not good enough. Advice has to be traceable to a policy, and a person has to stay accountable for the decision. That constrains the design from the start.

The relationship was never purely technical. Alongside the build, Humaneer drew on us for UX and product strategy advice, how the compliance intelligence should surface in the product, what to build next, how Mak should be positioned to the market. We advised throughout; every product and strategic decision stayed with Humaneer’s own team.

We built the pipeline that turns meeting transcripts, context, policies and user inputs into structured outputs: transcript processing, knowledge retrieval over the policy base, contextual reasoning, document generation and compliance guidance.

Trust and traceability drove the architecture. The system surfaces what it drew on, and it is designed to fit the workflow the product already had rather than asking users to adopt a new one.

The work has continued as Humaneer repositioned Mak from an HR chat assistant into what they call People Risk Intelligence: pulling meeting transcripts directly from Teams, Zoom and Google Meet, applying the same compliance intelligence Mak already used in chat to real conversations, and moving the dashboard from reporting what happened to recommending what to do about it. That repositioning work leaned heavily on our UX and product strategy input, sitting alongside the technical build rather than separate from it.

Conversation and context became practical next steps inside the existing product, with a human retaining sign-off, the point of the design rather than a limitation.

The work supported the company through VC fund pitches, its first customers, beta trials and early-adopter onboarding, alongside the feedback monitoring that told the team what was actually being used.

Mak now extends beyond chat: the same risk-spotting it applies to a typed question also runs across real meetings, so a manager gets the same protection without having to type anything up.