Work

Co-founder & Design Lead · 2026 · Reyda

An Agentic application for bid teams:
from decision to execution.

Reyda’s MVP helped teams find the right opportunities. I redesigned the product around what happened next: giving agents responsibility for the repetitive execution, and keeping judgement and accountability with the person.

The problem · At scale

Procurement decisions are
high-stakes and time-constrained.

Open on paper Tenders are standardised, consistently published, and open to any qualifying company across all member states.
Participation tells a different story Companies either drop out early or never complete the process. SMEs win most contracts by count, far less by value.
Digitised, without being simplified Teams still read, extract, and rebuild structure by hand. Manual interpretation repeats across every opportunity.
€2T EU public procurement About 14% of GDP
41.8% Single-bidder tenders Nearly doubled in a decade
<5% Cross-border participation Flat, year after year

V1

What I shipped first. Click through it.

Read the case study ↗

V1 · The impact

What V1 changed.

More tenders qualified In the same working hours
Hours instead of days To a confident go or no-go
Early funding secured Pilot data became the investment thesis

V1 · What it left behind

It worked. And it showed me the next opportunity.

The shortlist was full of tenders they could genuinely win, and everything after the yes was still typed by hand. In 2021, 41.8% of EU tender procedures drew a single bid, up from 23.5% a decade earlier, according to the European Court of Auditors. The winner is often whoever gets a complete bid in on time.

The opportunity

Teams were turning down tenders they could win.

Take the repetitive half off them.

Forty-three form fields retyped from documents they already hold, and the same evidence hunted down again for every bid.

Give the time back to the part only they can do.

Positioning, pricing, which win themes to lead with. The parts that decide whether you win, and the parts a person is actually for.

The new experience · Home

Home is a command centre.

I designed Home around one question: what does the bid manager need to know, decide or act on right now?

Instead of checking every bid, they get one view of what changed, what needs their judgement and what needs their next action. Agents keep the rest moving and surface only what they cannot finish, so nobody goes looking for work.

The new experience · Prepare

The Bid Room is where agents take over.

Once the bid manager decides to pursue, the agents research the requirements, find the evidence and fill in what Reyda already knows. They open the bid to work in progress, and pick up the gaps and the decisions that need a person.

The new experience · Bids monitor

Multiple bids, moving at once

Instead of working through bids sequentially, agents keep several opportunities progressing in parallel. The bid manager sees where each bid stands and steps in only when needed.

The new experience · Trust

Every answer can be interrogated.

Trust doesn’t come from asking the bid manager to accept an agent’s answer. It comes from letting them question it. They can see what Reyda used, where the evidence came from, and what’s still missing before deciding what to do.

Outcomes

What V2 changed, with one design partner.

Time to value 9 → 1 Days to a first draft ready for review
Tenders per person 1 → 4 Carried at once, with the repetitive half handled
Win rate 12% → 22% Nearly double, on the bids that went through Reyda* * We treated that as an observed business outcome, not a causal UX metric. The stronger product measures were time to first draft and number of bids carried per person.

V1 made finding the right tenders fast. V2 is about doing them: one design partner, first two quarters, on real tenders.

01 — 18