Every estimator knows the feeling of opening an RFP on a Monday, seeing a due date of Friday, and learning that three other firms have apparently been "in conversations" with the owner since spring. Nobody told you the party had started. You just got the invitation to the cleanup.

A startup called Cascade is built around exactly that gap, and it points at a better answer to a question GCs keep asking: how do you win more public bids?

The most reliable way to win more public bids is to stop discovering them when the RFP posts. Watch the public signals that precede a solicitation, such as bond measures, capital plans, permits and budget approvals, so you know a project is coming months ahead, then spend your estimating hours on the few pursuits where you have real fit and a relationship. AI is useful for the watching and sorting, while the decision to pursue and the final number stay with your people.

What did Cascade actually announce?

According to coverage of the July 21, 2026 announcement, New York-based Cascade raised $3.5 million in seed funding led by a16z Speedrun, with Ada Ventures, Blitzscaling Ventures, Indico Capital, shuckerVC, G2C Ventures and Snowball VC also participating. The product scans bond filings, permits, government capital plans and budget announcements to flag likely projects before a request for proposals goes out. Its customers are architecture, engineering and construction firms.

The company says customers have surfaced more than $10 billion in potential projects. Read that carefully: it describes identified opportunities, not revenue won. The one conversion example in the coverage is a $6 million project won by an architecture firm, and it comes from a customer testimonial. Cascade has not disclosed valuation, revenue or customer count, and coverage notes that pre-bid intelligence, private-market signals, bid leveling and team matching are still labeled "soon" on its product page. This is an early-stage bet, not a proven win-rate lever.

Why does finding the project early change win rate?

Win rate is a ratio, and most teams only work on the denominator once the RFP is in hand. Knowing about a school bond or a capital plan line item early lets you do things a five-day bid window never allows: meet the owner or the design team, line up the subs you will want, and decide in advance whether the job fits your backlog and bonding. A no-bid made in July costs nothing. A no-bid made on Thursday after two days of takeoff costs a lot.

That connects to a theme this blog has covered before: estimator hours are the scarce resource, so the cheapest win-rate improvement is usually bidding fewer of the wrong jobs. Early signals feed that decision with something better than gut feel.

Is this the same as bid intelligence on open solicitations?

Not quite, and the distinction matters when you shop. Infotech announced Bid Intelligence in April 2026 as an AI suite inside its BidX platform, aimed at helping civil contractors pick high-fit opportunities and reduce bid risk. That is about ranking solicitations that already exist. Cascade is aiming one step earlier, at the pipeline before solicitation. Both are useful, and neither replaces the estimator reading the documents and owning the price. I only saw these descriptions in search results and could not open the full pages, so check the vendors' own materials for specifics.

What should you ask before buying a pre-RFP tool?

Ask which public data sources it covers in your region, how often they refresh, and how it scores fit for your trades and bonding capacity. Ask for conversion data from customers you can call, not pipeline totals. And ask where a human reviews the output, because a confident score built on a stale capital plan is still a stale capital plan. Think of it as scouting, not scheduling: it tells you where to look, and the pursuit still depends on relationships and a competitive number.

The firms that win public work consistently have rarely been the ones with the fastest takeoff on bid day. They are the ones who were already in the room. If AI makes it cheap for a ten-person shop to hear the starting gun months earlier, that is a more interesting shift than another faster-counting-doors demo.

Frequently asked questions

How can a contractor win more public bids?

Start earlier and bid more selectively. Track bond measures, capital plans, permits and budget approvals so you know about a project months before the RFP posts, then pursue the few where you have real fit and a relationship. AI can watch those public signals at scale, but the decision to pursue and the final price stay with your team.

What is pre-RFP intelligence?

It is the practice of finding likely projects from public signals such as bond filings, permits, government capital plans and budget announcements before a formal request for proposals is published. It gives a contractor time to build relationships and prepare, instead of reacting to a short bid window.

Should I trust a vendor's claim about opportunities surfaced?

Treat it as pipeline, not revenue. Cascade, for example, says its customers have surfaced more than $10 billion in potential projects, but that measures identified opportunities, not contracts won. Ask any vendor for win-conversion data from customers you can speak to.

Sources