On a typical mid-rise construction project in Colombo, a quantity surveyor still spends days extracting line items from drawings, reconciling them against contractor submissions, and rebuilding the bill of quantities every time a wall moves. Concolabs, a pre-seed company headquartered in Western Province, Sri Lanka, is trying to compress that work into something closer to a software workflow. Founded by Ishini Saparamadu, the company sells what it describes as AI-driven Revit automation, intelligent document management, automated BOQ generation, construction ERP, and smart purchase order processing [Concolabs]. The bet is that construction, long resistant to vertical software, is finally ready to buy AI tools that speak its native language of drawings, specifications, and contracts.
The bet
Concolabs is anchoring its product line around the daily artifacts of a construction project rather than around general-purpose AI. Revit automation targets the building information modeling layer where architects and engineers spend their hours. The bill of quantities generator goes after the quantity surveyor's core workflow, the line-by-line accounting of every cubic meter of concrete and every meter of conduit.
Document management and purchase order processing wrap around procurement, where leakage and disputes are common. According to the company's site, these are sold as discrete tools that fit into existing project workflows rather than as a single suite that demands rip-and-replace adoption [Concolabs]. For a pre-seed company with ten employees [RocketReach], that wedge strategy is sensible: each module can be sold on its own, and each generates the data exhaustion that makes the next module easier to deploy.
The customer Concolabs is implicitly courting is the mid-market contractor or consultancy in South Asia and the Gulf, the kind of firm that runs dozens of projects in parallel without the IT budget of a Bechtel or a Larsen & Toubro. Saparamadu's own background as a Quantity Surveying graduate from the University of Moratuwa and a former staffer at International Construction Consortium and ITL [RocketReach] suggests the product is being designed by someone who has lived the workflow it automates.
Why it could be big
Global construction is one of the largest industries on earth and one of the least digitized. The category has produced a handful of large software outcomes (Procore, Autodesk Construction Cloud, Trimble) but vast swaths of the work, particularly outside North America and Western Europe, still happen in Excel, WhatsApp, and PDF. AI that can read a drawing set and produce a defensible BOQ is the kind of capability that, if it actually works at production quality, collapses a week of senior labor into an afternoon of review. That is the upside case Concolabs is reaching for.
The geography matters too. Sri Lanka, India, Bangladesh, and the Gulf states together represent an enormous pipeline of infrastructure and real estate spending, with a quantity surveying tradition that is rigorous and standardized. A tool built first for that market has a credible path into the British, Australian, and Middle Eastern markets that share the same QS conventions. Saparamadu has been recognized in the regional AI community for earlier tools called Roboclause, for contract interpretation, and MeasureonAir, for project management [aidatascientists.com], which suggests the BOQ and document modules are extensions of work she has been iterating on for years rather than a cold start.
The team and traction
Saparamadu is the founder and CEO [RocketReach]. Beyond her industry posts at ITL and International Construction Consortium, she has been a research assistant at the University of Moratuwa and an Enterprise AI Solutions Engineer at Intelligent Label Solutions, and graduated with a 4.02/4.2 GPA from Moratuwa's quantity surveying program [youngscientistawards.com]. She has also attended the Stanford Graduate School of Business [LinkedIn]. The early engineering bench includes Ameera Weerasuriya and Sajitha Jayawickrama, both software engineers [RocketReach]. Headcount stands at ten [RocketReach].
Employees | 10 | people
That is a tight team for a company taking on five product surfaces at once, which means the near-term roadmap likely depends on which module pulls early paid usage. Revit automation and BOQ generation are the most defensible candidates because they map to billable hours that contractors already pay for.
The honest counterfactual
The bear case is that construction AI is a crowded and difficult category. Autodesk has been embedding AI into its own BIM stack, Procore has been acquiring around document intelligence, and a wave of Y Combinator and seed-stage entrants have been chasing the BOQ and takeoff problem in North America. A pre-seed company in Colombo with no disclosed funding [Tracxn] is operating with less capital than incumbents spend on a single product manager. The bull answer is that none of those incumbents are designed around the South Asian and Gulf quantity surveying convention, where the BOQ is a contractually binding document with a structure unlike the takeoff sheets used in the United States. A vertical-within-a-vertical wedge, executed by a founder who has written BOQs herself, is exactly the kind of position that has historically produced durable software companies in markets the giants overlook. The Tracxn description of the work as "revolutionary" [Tracxn] is the kind of marketing language that should be discounted; the underlying product surface, BOQ automation tied to Revit, is concrete and measurable.
What to watch
The next twelve months will hinge on three things. First, whether Concolabs can convert pilots with Sri Lankan and regional contractors into paid recurring contracts, which would validate that the BOQ and Revit modules are accurate enough to trust on a live project. Second, whether the company raises a priced seed round; with ten employees and no disclosed external capital [Tracxn], a first institutional round would buy the runway to deepen the Revit integration and hire a sales lead for the Gulf. Third, whether Saparamadu can recruit a senior engineering hire to sit alongside her current team as the product surface expands. None of these milestones are guaranteed, but each is observable, and together they will tell the market whether construction AI built in Colombo can travel.
Pulse Raman covers the workflows that real industries actually run on. Standard of care today in mid-market construction QS: Excel, PDF markup, and a senior surveyor's judgment, billed by the hour.