Research, analyse, exclude, improve → all with a click of a button. A tender analysis system that spots opportunities for you.
Automate your processes and work 10x faster. Check out how this became possible for BRXPL with the Market Researcher system.
This system:
- Researches chosen public tender platforms
- Analyzes them based on the business intent
- Notifies about opportunities
- Assists with outreach based on user's decision
New Tender Alert
Score: 2/6 - High Relevance
Railway Dismantling Project - Steel Recovery
Est. value: €450,000 | Deadline: 5 days
Business context
Context
Scrap is sourced from various places. Apart from scrap selling companies a big sector is public procurement. Dismantling/Demolitions of power plants, rail tracks etc. are places where valuable raw material gets extracted.
Problem
Doing this manually is slow and expensive. There are tens of tenders platforms and thousands of tenders themselves. People cannot check enough data to have a high probability of spotting opportunities.
Consequence
Lost opportunity means that valuable material goes to your competition. You can't make profit without information on time. Being first equals $$. This app allows you to be the first to make the move in the market of tenders.
System at glance
This is an end-to-end automation that uses the power of AI to analyse the data while keeping the business intent in mind.
Phase 0
Discover
The system scrapes data from tender platforms.
Phase 1
Collect
The data from each scraped tender is collected and put into a queue for analysis.
Phase 2
Analyze
The system applies consistent decision rules to analyze the gathered data based on the user's needs.
Phase 3
Action
When opportunities are detected, the user gets notified.
It turns noisy portals into actionable data.
Deliverables: Value brought in
No manual checking. You do not have to spend any time controlling the system or staying up to date with platforms.
Every tender that you get notified about is business relevant. AI used in the app is prompted to evaluate the data according to your business's criteria and exclude all noise.
Every tender gets a score of relevance from 1 to 6. You are only notified about those that scored 1 to 3. That way you know straight away, which of the tenders are most worth looking at.
You receive the notifications in your chosen medium like email, Telegram or Whatsapp.
Decisions are made with a click of a button. You can choose which tenders should be ignored, handled by an assistant or by a human.
How this works
Data collection
The system monitors selected platforms via API connections or dedicated Playwright scrapers. After daily updates, tenders get deduplicated and saved in the database. This ensures reliable data for the rest of the process.
Interest Score
High Relevance
Interest Analysis
The content of each tender record is analyzed → all get a score of business relevance. That way the user will be notified about tenders that are worth looking at.
Notification & decision
User receives a structured notification about the relevant tenders. The user can either choose to handle the tender manually, or use an AI assistant to reach out to people responsible for the specific tender. Decision is done via buttons attached in every message.
Secure decision handling
Action buttons are signed and time limited until the tender loses its time worthiness, or closes. That way there is no risk of accidentally reaching out to people regarding closed tenders.
Wrong outreach is prevented. Unauthorized parties cannot send any messages on your behalf because links are encrypted and only authorized users connected to the app can make decisions.
Handling messy data
Not everything that the system receives is clear and well structured. Therefore correction logic is vital.
If contact data is unclear, or missing, the system verifies the record in a flow that performs corrections on the text and researches contact data of the contact person regarding a specific tender.
In case of incomplete data, the system will attempt to collect the tender information again to exclude possible errors.
Correction Loop
Verify → Correct → Re-try
Results and performance
| Aspect | Before | After |
|---|---|---|
| Time spent per 100 tenders checked | Manual browsing: 2h + | Automated filtering: decision making and reviewing found opportunities is the only manual activity. |
| Decision Speed | Hours/Days before the client even found the right tender after it has been announced | Everything is handled in minutes. |
| Consistency | Human judgement was faulty and inconsistent. Prone to errors. | Judgment is clearly defined and consistent every single day. |
Behind the scenes
Tech we've used:
Extensibility and adaptation
This solution works beyond scrap sourcing. The general logic is completely reusable and adaptable to any industry. What changes is the preferred/industry specific tender platforms and the business interest criteria that tenders will be filtered for.
Scrap
Construction
Energy
Transport
Want this for your business?
We can deploy a custom tender analysis system integrated to your workflow in less than 2 weeks.