Grew a one screen brief into a whole freight platform, from the first search to the closed deal, designed end to end and handed over finished.
The brief was a dashboard. What Fidelio actually needed was a whole product, one that sends a dispatcher's search to transporters, gathers every offer and conversation in one place and closes the deal, and none of it existed yet. I designed it end to end, as the only designer, and it was finished and handed over.
At a glanceOne search emails the right transporters for you, with templates, follow ups and nothing typed twice.
Read → 03 Every offer, one pipelineEach search becomes its own pipeline, with every reply, price and conversation side by side.
Read → 05 Built for a teamAgents, a manager who can step into their account, and notifications when a deal closes.
Read →Dispatchers found transporters by hand, and Fidelio wanted an MVP fast. The brief covered a dashboard, but every screen I designed raised the question the brief hadn't answered yet:
Answering them together with the founders turned a dashboard into a system. A lot of my job was mapping the depth nobody had planned for yet, before it surfaced in development.
A dispatcher describes the load once: the route, the vehicle, length, weight and loading dates. From there the system emails matching transporters on their behalf, with text it writes itself or a template of their own, and can follow up every few hours until someone answers.
Because the same routes and trucks come up again and again, I made sure nothing had to be typed twice: searches can be saved and reused from a row of recent routes, vehicles can be saved and appear first in the list, and email templates live in their own library.
Each search opens its own client pipeline. Replies land as rows: the company and its rating, the load, the price, and whether they've answered or are still waiting. Tabs split the pipeline into In Progress, Closed, Cancelled and Lost, so a dispatcher always knows what's still in play.
Opening a row slides in the whole conversation without leaving the list, with the deal's key facts pinned on top. From there it's one step to accept an offer or reply, written yourself or drafted by AI.
Closing a deal is where the MVP thinking was thinnest, so I designed what happens around it. Accepting one offer ends the search, which leaves every other offer hanging. So the moment a deal moves to Closed, a dialog explains that the search will go inactive and offers to move the rest to Lost, with a "don't ask me again" for people who already know.
After the job, the dispatcher rates the deal, and that rating becomes the company rating shown in every future pipeline. Agents can be rated too, and blacklisted, so over time the platform learns who is worth contacting.
Fidelio's dispatchers work as a team, so the product needed a manager's view as well. A manager sees every agent with their active and past searches, and can log in as any of them to step into their work. While they're in someone else's account, a banner says whose it is and links straight back, so nobody acts in the wrong account by accident.
Around that sit the pieces a team runs on: notifications when an agent closes a deal, and a documents list for CMRs, invoices and packing lists, showing who uploaded each one, who approved it and whether it was sent.
This was my first product with no existing system to lean on: no users to observe, no old screens to improve, only a brief and the questions it didn't ask. What I took from it is that on an MVP, design's biggest contribution often comes before the screens: following the flow to its end, so the gaps show up in a design review rather than in development.
The product was finished and handed over. Before it reached the market, the investors redirected the business, so it never had real users.
Designed solo as a freelance project, working directly with the founders and developers. No usage figures exist, as the product didn't launch.