From selling devices to a SaaS with its own channel

Emendu was a sales-led company with a SaaS on top, and the brief was not to manage a roadmap: it was to scale up a newly created SaaS within the company. The first step was validating who it was really for. To do that I joined the sales team in more than 20 meetings and interviewed 10 companies that were not in the pipeline.

That discovery produced a new ICP, and with it the whole 2025 roadmap: evolving the SaaS towards that customer's needs, reorienting LISA, our AI agent, and rewriting the product's language for the people who would actually use it.

The next step was iterating on how it was built. I drove the move from an external development agency to an in-house engineering team, and hiring the Tech Lead brought a technical roadmap with order and dates.

With the engineering team focused on the roadmap, I wanted to reach further: what internal stakeholders were asking for, I built myself with Claude Code. I created internal tools for Sales, for Account Managers and for the rest of the company. Not for me: each piece has its own team.

At Emendu, AI became part of the method. We reshaped LISA around SaaS management. Working prototypes became a key piece in how the product evolved, not only internally: they also let us put our ideas in customers' hands. And analysing interviews let me review not only the user but myself as an interviewer.

With the 2025 roadmap on the table, Sesame HR appeared, and it changed the nature of the project. Until then, optimising the digitisation of the operation was an efficiency project, but a partnership changes it at the root: the operation stops being only ours and has to work from inside someone else's product too.

Sesame HR is one of the largest HR software players in Spain, and the partnership brought Emendu's device management into its platform. What follows is how the work was done inside.

Internal and external discovery, to approach the operation with a new lens. Interviews with Operations, Sales and Account Managers, and with external sources too.

From there came a product spec with features and sub-processes; flow diagrams as code with their states and transitions, because a state diagram leaves no room for interpretation where a list of requirements does; and a navigable, low-fidelity HTML prototype, which became the only shared language of four audiences at once: engineering, Operations and Sales, the partner, and product.

The result: a win that few startups pull off.