Cutting tender qualification from weeks to hours

Co-founder & Design Lead · 2025 · Reyda

Cutting tender qualification
from weeks to hours.

Reducing cognitive load in document-heavy tendering through unified discovery, AI-powered comprehension, and guided qualification.

Timeline · Eight months

Eight months, from signal to funded.

I · Diagnose Signal + reframe

II · Research External + internal audit

III · Design Architecture + flows

IV · Land Pilot to deploy

M · 01 The signal Public procurement is a €2T market, yet most bid teams fall off after discovery — qualifying fit at speed is impossible.

M · 02 The reframe Discovery isn't the bottleneck. Comprehension and qualification are.

M · 03 24 conversations Six EU countries, four roles, one shared friction: teams burn days on tenders they were never going to win.

M · 04 Landscape audit Existing tools cover discovery and response. The middle — read, interpret, qualify — stays manual.

M · 05 Four principles Cut cognitive load · surface fit early · hold context across stages · make AI legible.

M · 06 Library as memory Past submissions, certifications, capability statements become a queryable structural layer.

M · 07 Bid Room Where the work happens. Memory + AI that reads and summarises each tender, every output checked against the source.

M · 08 Funded Pilot validated. 8× more tenders pursued at the same headcount. Investors signed in.

The impact

The result of this project.

More tenders qualified In the same working hours

Hours instead of weeks To a confident go or no-go

Early funding secured Pilot data became the investment thesis

The problem

Bid 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

Research & insights

Discovery wasn’t the bottleneck. Qualification was.

My co-founder and I started with a simple assumption: improving how teams discover and track tenders would increase participation. Two views into the work corrected it.

I worked closely with a mid-sized EU bid team, sitting in live bid meetings and shadowing active tenders, focused on the parts of the workflow that never appear in process documents: the re-reads, the backtracking. To test whether it held beyond one team, I spoke to 24 bid and tender professionals across six EU countries, public and private. The same patterns showed up: the friction concentrates at qualification, where teams interpret requirements and map them to internal capabilities. Most of the time is spent building enough clarity to decide.

Synthesis wall: the existing workflow mapped

To understand where to differentiate, I looked across the tools used in bid workflows. Most legacy products supported parts of the process, but the gaps were consistent: discovery handled by portals and alert systems, response supported by workflow tools and proposal generators. The breakdown sat in between, as understanding and qualification remained manual. Across countries the workflow stayed the same; only the surrounding infrastructure differed.

The users

Four roles, one underlying need.

From the 24 conversations, I mapped the people the product had to work for. Titles varied and responsibilities overlapped, but four roles kept appearing in every team.

Go / no-go decisions Tender manager I need to determine whether a tender is worth pursuing, so that I can avoid spending time on opportunities we will not win.

Requirements & viability Bid manager I need to understand key requirements without reading everything, so that I can assess viability within hours instead of days.

Gaps & rework Project lead I need to identify what is missing early, so that I can avoid late-stage blockers and rework.

Evaluation to response Bid team member I need to move from evaluation to response without starting from scratch, so that we can act quickly when an opportunity is a strong fit.

Storyboard · The current situation

Weeks from discovery to decision.

I put everything we had heard onto one storyboard — every step, owner, and duration of the existing workflow — so we could see where the time actually went.

3–7 daysDiscover tenders

ManualFilter & shortlist

HoursDownload documents

5–10+ daysReview hundreds
of pages

3–7 daysExtract requirements

3–5 daysCross-check capabilities

DaysAlign with stakeholders

MeetingDecide to proceed

5–15+ daysBegin response

Across teams the goal was consistent: move from discovery to submission with confidence.

User goals

Three goals, one product.

Mapping the story surfaced the discovery: underneath four roles sat three core goals. Everything the product does answers one of them.

User goal 01

As a tender manager, I need to decide whether a tender is worth pursuing in minutes, not days, so that we avoid spending time on opportunities we will not win.

User goal 02

As a tender manager, I need to understand key requirements without reading everything, so that I can assess viability within hours instead of days.

User goal 03

As a tender manager, I need to move from decision to submission without starting from scratch, so that we can act quickly when an opportunity is a strong fit.

Where we landed · The future

We redrafted what the future could look like.

The competitive analysis showed where the market wasn’t looking. The existing workflow showed where teams were quietly losing time. We took both, redrafted the future state — and went solving.

ContinuousTenders surfaced automatically

MinutesAI parses on upload

MinutesConversational research

InstantCompliance auto-mapped

SupportedGo/no-go decision

HoursProposal pre-filled

HoursReview & submit

Where we landed · The core problems

Three solves.

Solve 01

Making qualification faster

Qualification consumed most of the effort before proposal work could even begin. “Is this worth pursuing?” had to become a minutes-long question.

Solve 02

Turning documents into understanding

The bottleneck was interpretation, not access. Reduce the reading without reducing trust.

Solve 03

Turning understanding into action

Knowledge was recreated at every stage. One pipeline from decision to submission, context preserved.

The experience · It’s clickable

Try it. Every screen is live.

Watch: the product will walk itself through the solves. Or just take over and click.

Solve 01 · Making qualification faster · In detail

We designed around decisions, not documents: intent-based design.

Qualification ate most of the effort before proposal work began: days of monitoring portals, coordinating expertise and checking eligibility, all before the only question that mattered. “Is this worth pursuing?”

So we mapped what each role needed to decide sooner, against one principle: reduce the effort to qualify. Active work surfaces immediately, opportunities sit in one place, and signals arrive before the document review.

Outcome. What took 2 to 6+ weeks became a go/no-go within hours, in the first sitting.

Solve 02 · Turning documents into understanding · In detail

Reduce interpretation effort without reducing trust.

Even with a promising opportunity found, the slowest part was still ahead: requirements, obligations, risks and evaluation criteria buried in long documents. The bottleneck was not access to information. It was interpretation.

The goal was never to summarise. Early concepts leaned on AI output, but research kept returning to transparency and verification, so every extraction traces back to the original RFQ and the source stays one click away.

Outcome. Reading went from a multi-day slog to a conversation with the document. Teams stopped calling it reading and called it questioning.

Solve 03 · Turning understanding into action · In detail

The objective was not automation. It was continuity.

Research showed qualification work broke down the moment teams moved into compliance, forms, proposal writing and submission. Knowledge was recreated at every stage.

Now what qualification produces flows into everything downstream. Compliance board, forms and proposal stay together per tender, so qualification and submission are one connected experience, not two workflows.

Outcome. The same team, same hours, moved through 8× more tenders, and the pattern held across five more organisations of different sizes and sectors.

Validation · Testing the system in real workflows

A structured pilot — one company, then five.

Rather than ship broadly, I designed the rollout as a structured pilot. One company first, then five more, then a case for investors, then a direction the pilot itself surfaced.

Phase 01

Baseline and same-team pilot

I shadowed a mid-sized bid team over several weeks and tracked how long each workflow stage actually took. With working hours held constant, qualification moved from days of interpretation to hours of structured decisions.

Phase 02

Expanded across five more organisations

Once the first pilot stabilised, I extended the rollout to five more organisations of different sizes, sectors, and tender volumes. The pattern held: go/no-go within hours, drop-off on viable tenders fell sharply.

Phase 03

From pilot to investment thesis

Once the data was stable, I compiled the findings — workflow timings, the 8× outcome, behavioural patterns — into a case for investment, and presented it directly to prospective investors.

Phase 04

What the pilot surfaced next

Smaller organisations were qualified in part but not fully — close to winning on their own merits, reliably excluded by team size or thresholds. A direction we hadn’t originally scoped.

Reflection

AI was the premise the whole product was built around.

What I take most from this is how I now think about AI. The product worked because AI sat at the foundation, doing the heavy data work of reading, interpreting, and structuring that the workflow used to absorb on its own. Every other decision we made was built on top of that one, which made the rest of the system compose more cleanly.

And the only honest measure of whether the work succeeded is whether the people using it reached their goals faster and with more confidence. Output on its own says very little. The product earns its place when users move from hesitation to action.

V1 answered qualification. What it could not do was carry the work through drafting and submission, so we rebuilt it around agents.

Previous From product bet to App Store Scenra · 2026 Next Using an agentic system to solve food security at scale Oracle · 2024 to Present