A Calgary construction-software company is testing an AI agent that lets estimators query building drawings and specifications in plain language, then flag risks before they bid. PataBid’s Newton is in beta with a small customer group, with a commercial launch planned for 2027. The bet is that AI can help with a costly, detail-heavy job without getting the final say on the price.
Watch Desk analysis
What happened
PataBid CEO Melvin Newman says Newton is being developed alongside Quantify, the company’s estimating software. Contractors use Quantify to work with drawings and bid information; PataBid says more than 200 contractors use it and that over $2 billion in bid value has passed through the platform. The company announced a $2 million pre-seed round on 20 October to develop its AI, expand supplier partnerships and grow North American sales.
Newman says PataBid runs on hardware it owns in a Calgary data centre and trains models in-house. He says data stays within that system apart from an encrypted backup. Those are the company’s claims about its setup, not a general guarantee that keeping data in Canada alone settles questions about legal jurisdiction.
Why it matters
A construction estimate is not a low-stakes guess: missing a requirement in a drawing or contract can affect whether a contractor wins work and whether the job makes money. Newton is aimed at making project information easier to query and potential risks easier to spot, while leaving judgement with the estimator.
That division matters. Newman says experienced estimators carry knowledge that can be difficult to pass on, but the company also warns buyers not to treat AI output as unquestionable. In a field where a confident mistake can be expensive, “check the answer” is a sensible feature, not a lack of faith in the product.
Our read
This is a useful, specific test of AI in professional work: can a system help people find relevant details in dense documents without pretending it can make the business decision for them? PataBid’s claims about its platform and data controls are worth watching, but the key evidence will be how Newton performs on real bid work and what its users still need to verify. Estimators should be the ones checking the AI, not the other way round.
What to watch
- Whether Newton moves from a small beta group to its planned 2027 commercial launch.
- What the company can show about accuracy when querying drawings and specifications.
- Whether contractors find project risks they would otherwise have missed, and how they check those results.
Discussion spark: For construction estimates, should AI be allowed to flag risks and suggest figures, or should its role stop at finding information for an estimator to judge?
Sources and evidence
- Teaching AI to price a job you can’t afford to get wrong – Digital Journal (30 October 2026, 18:51 UTC)
Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.