Chasing Bids Manually.Then Winning Them Systematically.
2–3h → mins
per bid submission
An Alberta demolition and site-clearing contractor was limited not by capacity but by bandwidth. We built the system that watches every relevant procurement source, scores each opportunity against what the company can actually deliver, and hands the operator a priced submission to review.
The Constraint Was Never the Crew.
The contractor was spending hours every week manually checking government and municipal procurement portals. Opportunities were missed simply because nobody had time to look, and the ones that were caught meant writing a proposal from scratch against a deadline.
That is a bandwidth problem wearing the costume of a sales problem. The equipment and the crews were available; what was not available was the office time to find, qualify and price enough work to keep them busy.
So the build targeted the assembly, not the decision. The system finds and prices the opportunity; a person still decides whether to chase it and still signs off on what gets submitted.
At a glance
- Client
- A demolition company (anonymous)
- Industry
- Demolition & Site Clearing
- Location
- Alberta
- Build
- Bid intelligence & proposal system
- Result
- 2–3 hrs → minutes per bid
- Status
- Live in production · client-reported
Find It, Qualify It, Price It, Draft It.
Four stages that used to be one long manual job, each of which the system now does before an operator opens it.
Watch every source
Multiple Alberta government and municipal procurement sources monitored around the clock, so an opportunity is caught whether or not anyone had time to check that morning.
Score before anyone looks
Each posting is scored against the company's actual service lines, equipment, crew capacity and geographic range. Irrelevant work is filtered out before it reaches a person, so the inbox holds real opportunities only.
Price from the real rate card
Estimated cost and bid price are calculated from the company's own equipment rates and target margin, so the operator opens a bid with crew days, equipment mix and a recommended number already in front of them.
The System Prepares the Bid.A Person Decides and Signs.
Nothing is submitted automatically. The system does the finding, the qualifying, the pricing and the drafting — then stops. The operator makes the go or no-go call, edits the submission, and signs off before anything returns to the portal. The leverage is in the assembly, not in removing the judgement.
Architecture & The Hard Parts.
The real engineering — what the system is built on, and the non-obvious problems that had to be solved to make it hold up in production.
Portals That Were Never Meant To Be Watched
Public procurement sources each publish differently, with their own formats, categories and quirks, and none of them are designed to be consumed by another system. The work was normalising heterogeneous postings into one comparable record — and, just as importantly, making the ingestion loud when a source changes shape, because the dangerous failure here is silently seeing nothing and believing there is nothing to see.
Scoring That Can Be Argued With
A relevance score is only useful if you can see why it landed where it did. Service lines, equipment, crew capacity and travel radius are encoded as explicit criteria rather than a black-box judgement, so a bid that gets filtered out can be examined — and the rules corrected — instead of quietly costing the company work nobody knew about.
A Number You Can Defend
The recommended price is computed from the company's own equipment rates, crew-day assumptions and target margin, not a generic model. That matters twice: the operator can sanity-check it in seconds because the inputs are theirs, and if the number is ever questioned after the fact, the basis of it is recoverable.
Drafting That Stops at Review
A complete submission is generated and staged, not sent. The operator edits and approves, and only then does the system handle the return trip to the portal. Submitting a wrong number to a public procurement body is not a recoverable error, which is exactly why the human gate sits where it does.
What It Deliberately Does Not Do.
- It does not decide which work to chase. Go or no-go is the operator's call, every time.
- It does not submit anything a person has not approved and signed off.
- It is not a general procurement database — it is scoped to the sources and service lines that are relevant to this contractor.
- It does not replace estimating judgement on unusual scopes; it removes the assembly work around the ordinary ones.
Out of respect for our client
We will talk architecture, but the specifics stay private: how this system prices, the rates it runs on, and how it is tuned to this business belong to our client, and we keep them that way. When we build for you, your systems and your numbers get the same discretion.
Your operation is next
What Is Your Two-Hour Bid?
30 minutes. We map how work gets found, qualified and priced in your operation today, and show you what it would take to build the fix — whether you build it with us or not.
Edmonton, Alberta · No commitment · 587-937-6948