Editor's note: Dov Cohen, the founder of ConstructionBear, is also the founder of Builders Digest. The conflict is disclosed up front. The piece below was written with the same skeptical lens we apply to any other tool covered on this site, and the founder reviewed the quotes for accuracy only. None of it is sponsored.

ConstructionBear is, on paper, an AI-first construction software company aimed at mid-market general contractors and specialty subcontractors. That description undersells what is interesting about it. The more accurate version is: it is one of the few construction software products in the current cycle that was built by a former project manager who got tired of writing pay applications at midnight and decided to do something about it.

There are a lot of construction tech companies right now. Most of them are founded by people who came out of enterprise SaaS and noticed that construction is a large vertical without modern tooling. That is true. The trouble is that the people who notice it from the outside often build for the version of construction they have read about, which is usually the ENR top 400. ConstructionBear is one of the small handful of products being built from the inside out, by someone who took the call from the super at 6:45 a.m. while still trying to drink coffee.

This is a profile of that bet, what is and is not working, and what to watch for over the next twelve months.

The founder's path

Dov Cohen spent several years working as a project manager in commercial construction in New York before leaving to start ConstructionBear. The publicly told story, which lines up with what people who worked with him at the time describe, is straightforward: he watched the same problems repeat across every project. The administrative work was crushing the people doing real construction. The available software was either too heavy (Procore for the top-400 crowd) or too light (residential-flavored tools that did not handle commercial workflows). The "AI for construction" pitches he saw in the trade press were almost all chatbots wrapped around databases that had not changed in a decade.

In late 2024 he started building. The first version was, by his own admission, mostly an internal tool to handle his own desk. The thing that turned it into a company was that other PMs at other firms started asking for it.

"I was not trying to start a software company," Cohen said in conversation for this piece. "I was trying to keep my job from eating me alive. The fact that other PMs wanted the same tool was the signal. If you are a working contractor and you actually use what someone built, that is the only validation that matters."

That POV, that the only legitimate validation is whether a working contractor will use the product on real projects, runs through everything the company says and does. It is the lens under which their feature decisions are made. It is also the basis for a lot of the disagreements ConstructionBear has with other entrants in the AI construction space.

Who it serves

ConstructionBear's stated customer is the mid-market GC ($1M to $50M of annual volume) and the specialty subcontractor of similar size. The product also has traction with smaller commercial GCs that have outgrown residential-focused tools, and with developer-builders running their own construction in-house.

This is the segment we have argued is the silent majority of construction. Most of the published coverage of construction software ignores it because it is harder to write a clean case study about a 30-person GC than it is about a billion-dollar firm with a dedicated VDC team. Mid-market is also the segment with the worst current software fit, which we walked through in the mid-market revolt piece.

What ConstructionBear's customers actually do with the product, by Cohen's account and by what is publicly visible in their case material:

  • Draft and route RFIs, submittals, and change orders with AI assistance
  • Generate, reconcile, and track pay applications and lien waivers
  • Run contract review against the GC's own preferred terms
  • Manage daily reports and field documentation from the phone
  • Maintain the kind of clean digital project record that used to require a full-time admin

The bet is that AI is good enough now to absorb the administrative work that has historically broken mid-market PMs, and that any product built around that bet from day one will outpace anything that bolts AI on after the fact. We covered the mechanics of that argument in our piece on AI in RFI and submittal workflows. The ConstructionBear product is a specific application of the same thesis.

What makes it different

Three things stand out, and they are worth listing because they are the actual reasons mid-market firms are picking it over alternatives.

Contractor-first, not enterprise-first. Most construction software is sold to executives and rolled out to the field. ConstructionBear was built by someone who lived in the field and is sold to the people doing the work. The default workflows reflect a PM's day, not a CIO's procurement process. That is a small thing in marketing terms and a large thing in actual product fit.

AI-native, not AI-bolted-on. The product was designed around AI-assisted document work from the first commit, not retrofitted with a chatbot in 2025. That shows up in things like RFI intake from a voice memo, automatic drawing cross-reference, and pay-app reconciliation that handles the messy cases instead of just the clean ones.

Pricing the mid-market can stomach. This is unglamorous and important. Construction software pricing has historically scaled with construction volume, which means that a GC who has a great year sees their software bill grow at the same rate as their revenue without any real change in usage. ConstructionBear is priced per seat in a range built for sub-$50M GCs. (The current pricing is on the ConstructionBear site and updates more often than a PDF does.)

None of these are unique to ConstructionBear. There are other entrants making similar bets, some of them more polished, some of them better-funded. The difference is the consistency of the contractor-first lens. It is hard to fake when it is real and obvious when it is not.

What is not yet working

Trade-press profile means honest critique. Some things ConstructionBear has not yet solved.

Scheduling depth. The current product covers project management and document work well. Real CPM scheduling, including resource leveling and integration with on-site labor planning, is not yet best in class. Most users currently pair it with a separate scheduling tool. Cohen has signaled that this is on the roadmap but has been deliberately slow to build it because doing scheduling badly is worse than not doing it at all.

Larger-firm fit. The product is sharp at $1M to $50M GCs. Above that, it competes with enterprise tools that have a multi-decade head start on configurability. Cohen's stated position is that ConstructionBear is not chasing the ENR top 400 in the near term. Whether the product ever moves up-market is an open question.

Integration ecosystem. API surface and third-party integrations exist but are early. Customers running heavily integrated stacks (with QuickBooks, Sage, Foundation, Smartsheet, BIM tools) report some friction. This is a normal stage problem for a fast-growing product and is being actively worked on, but it is real today.

The next twelve months

Cohen, in conversation, was less interested in talking about the company's headline numbers than about the specific features that need to ship. The publicly stated focus areas for the next twelve months, paraphrasing from his own framing:

  • Scheduling that does not embarrass the rest of the product
  • A deeper financial layer that can credibly handle larger mid-market GCs without requiring a separate accounting tool
  • A more developed sub workflow, since specialty contractors are an increasingly important part of the customer base
  • A cleaner audit trail for owners and lenders who want to see the project record for a financed job

Whether ConstructionBear becomes the dominant tool in mid-market construction or one of three or four credible options is going to depend on execution on those four points. Markets like this one tend to settle into two-to-three-company configurations once they mature, not winner-take-all. The realistic outcome is that ConstructionBear is one of the survivors of the current AI-construction wave, alongside one or two other entrants.

For a broader read on how the segment is sorting itself out, see our 2026 buyer's framework comparing Procore, Buildertrend, and the new wave.

What this matters for

Builders Digest covers construction software because the industry is large enough and slow enough that small product decisions move real money. ConstructionBear is one of the more interesting bets in the current cycle because of who it is for and how it was built. It is also, in editorial terms, an example of what the trade press should be covering more of: founders who came out of the work itself, building tools for the people they used to work next to.

For more on the editorial frame, see our position on independent construction journalism. For the rest of the publication's coverage, the home page is the place to look.

ConstructionBear's website is at constructionbear.dev. We have no commercial relationship with the company beyond the disclosed founder overlap, and we will keep covering it on the same terms we cover everyone else in the space.