EditalPro
The public-tender radar that finds the right opportunity before the competition: multi-portal monitoring with AI scoring, from notice to proposal.
Too many notices, too little time.
Brazil publishes thousands of tenders a day, scattered across dozens of portals: PNCP, BEC-SP, state and municipal sites, each with its own format. For a company that sells to the public sector, finding the right notice is slow, manual, expensive prospecting.
Those who arrive late miss the deadline. Those who try to cover everything drown the team in noise. What was missing was a radar that filtered what matters and showed, clearly, where it was worth betting.
Opportunity feed with profile match and AI relevance scoring.
From zero to live product.
Discovery with the people who live the pain
I talked to sales and bidding teams to map the real flow: where they search, what they discard, why they miss deadlines. The core insight: the problem isn't a lack of notices, it's excess without prioritization.
Data ingestion & normalization
I built connectors for the public portals (PNCP, BEC-SP and others), normalizing different formats into a single tender model (object, value, deadline, agency and location) ready to be filtered and scored.
AI relevance scoring
The heart of the product: a model that reads each notice and assigns a relevance score for the company's profile. Plus a swipe calibration: with each user "yes/no", the AI learns and refines the ranking. The more it's used, the sharper it gets.
Workflow through to proposal
Relevant notices become work: a proposals kanban, deadline alerts, and an opportunity map by state to see geographic concentration. From alert to submission, all in one place.
Go-to-market, solo
Landing, onboarding, billing and the first customers, all driven by me. Well-applied AI let me deliver like a team, from design to deploy, at one person's pace.
Proposals kanban: from triage to won, with value per stage.
The pieces of the radar.
Multi-portal monitoring
Continuous tender collection from PNCP, BEC-SP and other portals, in a single normalized feed.
AI score
Each notice gets a relevance score for the company profile: what matters rises to the top.
Swipe calibration
The user trains the model with a gesture. With each decision, the ranking gets more precise.
Opportunity map
A by-state view of tender concentration, to decide where to focus commercial effort.
Proposals kanban
From "interesting" to "submitted": each opportunity moves through a visual flow with owners and status.
Deadline alerts
Notifications of what's due soon, so no good opportunity dies on the calendar.
Opportunity map by state and municipality.
It wasn't just another tender aggregator. It was a radar that learns, and gives the team back time to sell.
PNCP · BEC-SP
product · code · GTM
with every swipe
notice to proposal