Why I Built Specchec (and Why I Was Up at 2 AM)

A few years back, I took a position as the lead estimator at a commercial millwork shop. On paper, my job was to bid work and run my team. In reality, I was spending 10 to 20 hours a week just deciding which projects were even worth bidding. Picking through plans and specs to figure out if they fit our workflows, our capabilities, our schedule. By the time I finished qualifying leads, I had no time left to actually estimate them. Which is how I ended up at 2 AM once or twice a week, slouched over my computer in Bluebeam, trying to crank out a competitive number on three hours of sleep. You probably know the feeling. I tried everything to dig out. Checklists. Smart templates. Then ChatGPT. That had to be the answer, right? It wasn't. Generic AI tools weren't built for this work. They forgot instructions halfway through a 400-page manual. They hit daily limits in the middle of a 6 PM deadline. They handed back "specifications" that were just confidently wrong. You can't bid a job off a confidently wrong document. You'll lose it. Or worse, win it. What I figured out is that no single tool gets you there. Not AI alone. Not a checklist alone. Not a sharp human alone, working on broken sleep. But software, AI, and a tightly defined task, all working in the same lane. That combination actually makes sense of the mess architects and owners hand us. That's Specchec. It does one thing, and it does it well: it reads the specs and drawings, and it pulls out exactly what a casework or millwork estimator needs to make a bid/no-bid decision and to start estimating. Bid submission requirements. Cabinet specs. Manufacturers. PLAM grades, colors, and finishes. Countertops. Hinges, pulls, slides. Materials legend. Prevailing wage and Davis-Bacon flags. Addenda tracked in order. RFI recommendations when the docs contradict themselves, and they always do. A 400-page manual turns into a clean PDF in under 10 minutes. The 30-minute hand-read on every spec is gone. The "is this even worth bidding" question gets answered before lunch instead of at midnight. I want to be clear about one thing: Specchec is not trying to replace estimators. It can't. Bid/no-bid judgment, relationships, scope strategy, pricing instincts. That's the job, and that's what experienced humans bring. Specchec just takes the brutal, hours-long document-reading part off your plate so you can do the part only you can do. If you're an estimator running yourself into the ground every bid week, give it a shot. Two free reports, no card. If it saves you the night I was trying to save when I built it, that's the win. I wish you the best and hope you get as much out of this as I have. — Morgan Sangre Strategies / Specchec