The Saturday problem
It's 9pm on a Saturday. A contractor we know — let's call him Marcus — is at the kitchen table with his laptop, a yellow legal pad, and four browser tabs open: RSMeans, Home Depot, his supplier's PDF catalog from 2022, and a spreadsheet he built in 2018 that he's scared to touch.
He has a bid due Monday. It's a $90,000 bathroom renovation. The margin on a job like this can swing 12 points depending on how good his estimate is.
Marcus is a brilliant contractor. He's built hundreds of beautiful spaces. He has 14 years of trade knowledge in his head that no software can fully replicate. But he's spending his Saturday night in Excel.
We thought he deserved better.
Why estimating is hard
Estimating isn't just math. It's knowledge work. To price a bathroom renovation correctly, you need to know:
- What materials cost in this region, this month (not national averages)
- What the labor rate is for your specific subs and crew
- Which items have long lead times that could blow up your schedule
- What the local code requires for this type of work
- What your markup should be given your current backlog
Excel doesn't know any of that. The big legacy estimating tools know some of it, but they're expensive, slow, and feel like they were designed by people who've never walked a job site.
The existing tools aren't the answer
We looked at what contractors were using. The options are:
Spreadsheets: Fast to start, painful to maintain, zero intelligence. Every new job means starting from scratch or copying a sheet that's full of last year's prices.
Legacy estimating software: Built in the 90s, priced at $300-500/month, requires a week of training. For most contractors, the cost and friction isn't worth it.
General construction management tools: Great for project management, not for estimating. They handle what happens after you win the bid — not how you win it.
None of these tools help you win the bid faster. None of them learn from the jobs you do. None of them help you show a client what their project will look like.
What changed: AI that understands trades
The last two years of AI development gave us something we didn't have before: models that can understand the semantic complexity of trade work.
Not just "kitchen renovation" as a category, but "remove and replace 22 linear feet of cabinets, add an island, move the sink 6 feet, upgrade to quartzite counters, and add under-cabinet lighting" — and generate a reasonable line-by-line estimate with materials, labor hours, and regional pricing.
We spent a year building and testing trade-specific AI agents. A kitchen specialist. A bath specialist. Flooring, electrical, plumbing, framing, exterior, decking. Each trained on the specifics of its trade. Each generating estimates a contractor can actually use, not just a ballpark.
The key insight was that generalist AI makes poor estimators. You need specialists who understand the difference between a linear foot of crown molding and a linear foot of baseboard, who know that 12x24 tile requires a different substrate than 4x4, who understand that "rewire the panel" is a permit-required job in most jurisdictions.
The visualization breakthrough
The other thing that changed: image generation got good enough to be useful.
Contractors have always struggled with one part of the client relationship: helping clients visualize what they're paying for. "Trust me, it'll look great" isn't a sales strategy.
We built a visualization engine that takes a real photo of the existing space and generates a photorealistic "after" image. Not a render, not a sketch — a realistic picture of what the finished space could look like with your proposed materials and finishes.
More importantly, we built object detection that works backward: it scans the visualization for cabinets, counters, fixtures, flooring, lighting, and trim, and turns each detected object into an estimate line item. The same image you show the client to close the deal becomes the foundation of your estimate.
What Bidmate is, and what it isn't
Bidmate is not a replacement for a good estimator. A good estimator's judgment — knowing when a job is going to run over, knowing which clients are going to be difficult, knowing when to walk away from a project — that's hard-earned knowledge that no software can replicate.
What Bidmate does is give that estimator their Saturdays back.
It handles the parts of estimating that are mechanical: pulling materials, pricing them against your region, checking lead times, building the line items, formatting the document, sending it to the client, getting it signed.
The estimator's job is to apply their judgment to the output, make the adjustments that only experience can make, and get the bid out the door in 45 minutes instead of 4 hours.
What's next
We're a small team. We shipped fast and we're learning from every contractor who uses Bidmate in production. The roadmap is driven entirely by what makes contractors more likely to win their next bid.
If you're a contractor who's tired of Saturday estimating sessions, we built this for you. It's free to start. We'd love to know what you think.
— The Bidmate Team