Announcing Our Investment in Surfaice
by Nick Durham, Shadow Ventures · December 3, 2024
Originally published on Shadow Ventures
Shadow Ventures announced its investment in Surfaice on 3 December 2024, in a post written by partner Nick Durham. The summary below is ours; the announcement itself is Shadow's and is linked above and below.
The thesis
The argument starts with the physical footprint. Brands in retail, quick service and EV charging grow by building the same thing many times over — a store, a restaurant, a charging bay — each following a precise playbook built for speed and low risk. Durham's account of what that looks like in practice is unflattering: site approvals dragging on for months, long checklists, coordination handled by people, and workflows that never left the previous era.
Surfaice was described as the first startup applying agentic AI to that category of work, aimed squarely at site approval as the largest bottleneck to scale.
Why repeatable construction specifically
The post makes a data argument rather than a market-size one, and it is the part most worth reading in the original. Companies building simple structures at scale generate structured records as a by-product — playbooks, checklists, standard designs. The wider architecture, engineering and construction industry does not: its data is fragmented and unstructured, and models built on that footing make more mistakes. In engineering and construction, those mistakes are expensive.
That is the reason given for starting where the data is already clean.
What convinced them
Durham points to a demo: agents managing Gantt charts, placing RFI calls, updating schedules, and running contractor qualification and bidding. He frames the wave in terms of what came before it — the last real shift in project management was the move to software and mobile, with Procore and PlanGrid as the landmarks, and agentic AI as the one following it.
The post also raises the objection directly rather than around it. A customer had worried that agents would produce inaccurate results and need checking every time. The answer given is the pairing this site argues everywhere else: work in industries where the data is structured, and build human-in-the-loop systems on the customer's own playbooks, leaving anything that cannot be automated accurately under human verification.
The team named
The announcement credits three people — Alim Uderbekov, Slava Solonitsyn, founder of Mighty Buildings, and Omar Zhan — citing their experience across construction marketplaces, prefab systems and project management at scale. It closes on the partnership itself: “We’re excited to partner with them.”
Read the announcement
The full post, in Durham's own words, is on Shadow Ventures: Announcing Our Investment in Surfaice.