Stoa
AI software for construction sites that consolidates project documents and identifies mistakes early.
Website: https://stoabuild.com/
Cover Block
From the public record
| Field | Value |
|---|---|
| Name | Stoa |
| Tagline | AI software for construction sites that consolidates project documents and identifies mistakes early [Sesamers, November 2026] |
| Headquarters | Helsinki, Finland [Startup Researcher, November 2026] |
| Founded | 2026 [Startup Researcher, November 2026] |
| Stage | Pre-Seed [Startup Researcher, November 2026] |
| Business model | SaaS |
| Industry | Proptech |
| Technology | AI / Machine Learning |
| Geography | Western Europe |
| Growth profile | Venture Scale |
| Founding team | Co-Founders (3+) [Today’s Startup News, November 2026] |
| Funding label | Pre-seed [Startup Researcher, November 2026] |
| Total disclosed | $3,700,000 [Startup Researcher, November 2026] |
Links
From the public record
- Website: https://stoabuild.com/
The Short Version
PUBLIC Stoa builds AI software for construction sites that pulls together project documents and answers operational questions against those records, a proposition that merits attention now because the company paired that product pitch with a newly announced €3.5 million pre-seed round and reported early deployment across more than 25 Nordic sites [Sesamers, November 2026] [Startup Researcher, November 2026]. The immediate appeal is straightforward: construction remains document-heavy and coordination-intensive, and Stoa is positioning itself as a way to surface mistakes earlier by unifying drawings, schedules, quality records, and safety reports into one queryable layer [Sesamers, November 2026] [The AI Landscape, November 2026].
The founding story is still early and only partially documented in public, but the company is presented as a Helsinki-based team formed around software, AI, and construction experience, with Johannes Korpela identified as CEO alongside co-founders Taavi Paananen, Camilla Komulainen, Eino Miettinen, and Pyry Pohjanoksa [Finsmes, November 2026] [LinkedIn, retrieved 2026]. Public reporting also ties the broader team to backgrounds at Supercell, AWS, SRV, and Peikko Group, which is relevant because the product sits at the intersection of enterprise software and on-site construction workflows [Sesamers, November 2026] [Today’s Startup News, November 2026].
On product, the clearest point of differentiation is not general-purpose AI but retrieval over project-specific records with citations back to source documents, a design choice that fits a category where auditability matters more than conversational polish [stoabuild.com, retrieved 2026] [Pomegra, November 2026]. Company materials say the platform reads drawings, specifications, contracts, schedules, and site records; connects with systems such as M-Files, SharePoint, Dalux, and Congrid; preserves existing permissions; and does not train models on customer project data, though those claims remain primarily company-sourced [stoabuild.com, retrieved 2026].
The financing is modest in size but notable in composition. Stoa announced a €3.5 million pre-seed in November 2026 led by Lifeline Ventures, with angel backing that includes Pekka Lundmark, Alexander Ehrnrooth, Tuomas Kokkila, Tracy Young, and Ralph Gootee, and the business model is presented as SaaS for construction and data-center project teams [Startup Researcher, November 2026] [Sesamers, November 2026] [Today’s Startup News, November 2026].
Over the next 12 to 18 months, the public case will turn less on the funding headline and more on evidence of repeatable deployment, named customer validation, and proof that early site activity converts into durable usage inside large construction organizations [Sesamers, November 2026]. The current signals are encouraging but still thin: more than 25 Nordic sites is a useful starting point, while larger claims around €4 billion-plus of project value and time saved remain company-led and need independent corroboration before carrying much weight in an underwriting model [Sesamers, November 2026] [stoabuild.com, retrieved 2026] [CASRAI, 2026].
Unconfirmed -- This section mixes independently reported funding, team, and deployment facts with several material product and traction claims that are currently company-sourced.
Taxonomy Snapshot
| Axis | Value |
|---|---|
| Stage | Pre-Seed |
| Business Model | SaaS |
| Industry / Vertical | Proptech |
| Technology Type | AI / Machine Learning |
| Geography | Western Europe |
| Growth Profile | Venture Scale |
| Founding Team | Co-Founders (3+) |
| Funding | Pre-seed, total disclosed approximately $3.7M |
The Company in Brief
From the public record The basic company facts are thin but unusually time-specific. Stoa presents itself as a Helsinki-based company building AI software for construction projects, with a stated focus on answering questions from project documents and surfacing mistakes earlier [stoabuild.com]. The company is associated with the domain stoabuild.com, which matters here because there are multiple businesses using the Stoa name and the website is the clearest public identifier for this construction-focused entity [stoabuild.com].
The public record also points to a very recent formation. Startup Researcher reported in November 2026 that Stoa had raised a €3.5 million pre-seed round led by Lifeline Ventures, and the company website published the same financing announcement, describing backing from a mix of venture investors and construction-linked operators [Startup Researcher, November 2026] [stoabuild.com]. A five-person founding team is named in public materials and secondary coverage, but this section stays with the website-backed company identity rather than assigning legal-entity detail that is not established in the cited sources [stoabuild.com] [Startup Researcher, November 2026].
Chronologically, the first clear public milestone is the November 2026 pre-seed announcement. By that point, the company website said the product was already in use on construction projects and described integrations with document systems such as M-Files, SharePoint, Dalux, and Congrid, positioning the product as a layer over existing project records rather than a replacement workflow [stoabuild.com]. The same website also states that customer data is hosted in the EU and not used to train the company’s AI models, which helps define the early operating posture even if broader corporate history remains sparse [stoabuild.com].
Unconfirmed -- This section relies primarily on the company website for identity, headquarters, and product-level company description, with partial corroboration from Startup Researcher on the November 2026 financing event [stoabuild.com] [Startup Researcher, November 2026].
What They Have Built
Mixed sourcing
Stoa is selling into one of construction's least elegant problems: too many critical decisions still depend on someone finding the right drawing set, schedule revision, inspection note, or contract clause quickly enough to matter. Public product descriptions are consistent on the core workflow. The platform consolidates project documents, including drawings, schedules, quality records, and safety reports, then lets teams ask questions about a project's current state and surface mistakes earlier [Sesamers, November 2026] [Startup Researcher, November 2026]. The company's website goes a step further on how that interaction works, stating that the system reads drawings, specifications, contracts, schedules, and site records, and returns answers with the underlying document and page for each response [stoabuild.com, retrieved 2026].
The implementation details that are public point to a retrieval layer built around existing enterprise systems rather than a rip-and-replace workflow. According to the company site, Stoa connects to tools such as M-Files, SharePoint, Dalux, and Congrid with read-only sync while preserving existing permissions [stoabuild.com, retrieved 2026]. Its terms state that the service reads documents from connected systems and direct uploads but does not modify documents in those systems [Terms of Service - Stoa, retrieved 2026], and its security materials say customer project data is not used to train AI models, answers are grounded only in the customer's own project data, and the service is hosted in the EU with EU AI Act transparency commitments [Security | Stoa, retrieved 2026]. Those claims fit the category's current buying criteria, especially for document-heavy projects such as data centers, but at this stage most of the detailed architecture and security posture remains company-described rather than independently tested in public reporting [Finsmes, November 2026].
What can be said with some confidence is that Stoa's product positioning is narrower than a general-purpose chatbot and broader than a document repository. Press coverage and the company site both frame it as an AI operating layer for active construction projects, with emphasis on traceable answers and earlier error detection rather than design generation or field robotics [Sesamers, November 2026] [Pomegra, November 2026] [stoabuild.com, retrieved 2026]. That distinction matters because the product's credibility appears to rest on citation, permissions, and system connectivity, not on model novelty alone.
Unconfirmed -- Core product workflow is corroborated by Startup Researcher and Sesamers, but most technical, security, and integration specifics in this section come from company materials only.
Market Size and Demand
PUBLIC The market matters now because complex construction projects are producing more documents, more compliance burden, and more coordination risk at the same moment that general-purpose AI has made retrieval and summarization usable at the jobsite edge, at least in principle [Sesamers, November 2026] [Startup Researcher, November 2026].
Public evidence on Stoa's addressable market is still thin, so the cleanest way to frame demand is through the workflow the company says it is targeting: construction and data-center projects where drawings, schedules, quality records, safety reports, contracts, and site records sit across multiple systems and need to be queried against the current state of the project [Sesamers, November 2026] [stoabuild.com]. That is not a niche paperwork problem. It sits inside the broader construction software stack, and more specifically inside document control, field collaboration, QA and safety, and project intelligence tools that try to reduce rework and coordination errors. Stoa's own early positioning, particularly its emphasis on data centers and other complex sites, suggests a wedge into projects where the cost of a missed issue is high and the document set changes daily [Startup Researcher, November 2026] [Finsmes, November 2026].
Without a cited third-party TAM for this exact category, the market has to be treated as an analogous one rather than a measured Stoa-specific segment. The closest public analogue in the source set is construction technology software used by project owners, general contractors, and specialty teams to manage project information, field workflows, and compliance records. A second adjacent analogue is software tied to data-center development, where schedule slippage and change-order risk can carry unusually high downstream costs, although the source set does not provide a quantified market size for that segment [Startup Researcher, November 2026] [Sesamers, November 2026]. The practical read is that Stoa is selling into an existing budget line, software for project information and execution, while trying to create a higher-value layer around question answering and early error identification.
| Market lens | What the sources support | Relevance to Stoa |
|---|---|---|
| Core workflow market | AI software for construction sites that consolidates drawings, schedules, quality records, and safety reports [Sesamers, November 2026] | Directly aligned with stated product scope |
| Adjacent vertical | Data-center projects are named as an early focus [Startup Researcher, November 2026] [Finsmes, November 2026] | Suggests entry via high-complexity builds |
| Budget substitute | Existing systems such as M-Files, SharePoint, Dalux, and Congrid already hold project records [stoabuild.com] | Implies Stoa may layer on top of incumbent systems rather than replace them outright |
The table points to a familiar pattern in enterprise software: a new application category emerging on top of entrenched systems of record. If that framing is right, adoption may depend less on persuading firms to rip out legacy tooling and more on proving that an AI layer can sit safely above existing repositories and permissions [stoabuild.com].
Demand drivers in the public record are concrete, even if still early. Stoa is described as consolidating project documents and answering questions about a project's current state, with the stated goal of identifying mistakes earlier [Sesamers, November 2026]. The company also emphasizes source-grounded answers, read-only integrations into systems such as M-Files, SharePoint, Dalux, and Congrid, and preservation of existing permissions, which speaks directly to two common frictions in construction software adoption: fragmented records and reluctance to move critical documents into a new operational system [stoabuild.com] [Terms of Service - Stoa]. Reported deployment on more than 25 Nordic sites by November 2026 is not enough to establish category leadership, but it does indicate that the underlying problem is live rather than theoretical [Sesamers, November 2026].
The substitute markets are also clear. A project team can continue to operate through incumbent document repositories, field management tools, manual search across PDFs, and human coordination across site managers, engineers, and subcontractors. That is the status quo Stoa appears to be challenging [stoabuild.com] [Sesamers, November 2026]. The company's positioning is stronger if buyers view it as an intelligence layer that shortens decision time without requiring system replacement. It is weaker if procurement treats AI retrieval as a feature that incumbent project platforms can add natively over time.
Regulatory and macro forces cut both ways. On the positive side, Stoa's public security materials stress EU hosting, source-grounded answers from customer data only, no training on customer project data, and EU AI Act transparency, all of which are well chosen for a European construction buyer that will care about data handling and auditability [Security | Stoa] [stoabuild.com]. On the limiting side, construction remains cyclical, procurement-heavy, and often conservative about workflow changes. That means macro slowdowns in commercial building can lengthen sales cycles even when the product solves a real coordination problem, while regulation can help only if the compliance posture is strong enough to accelerate vendor approval rather than merely clear a minimum bar [Security | Stoa] [Startup Researcher, November 2026].
Unconfirmed -- This section relies primarily on company materials and startup press, with no independent third-party market sizing report in the source set and several material claims supported only by company disclosures.
Who Else Is Fighting for This
MIXED Stoa appears to be positioning itself less as a point tool for construction documentation and more as an AI layer that sits across the systems a site team already uses, which is a meaningful distinction in a market where workflow ownership often matters more than model quality [Sesamers, November 2026] [stoabuild.com, retrieved 2026].
The competitive map is still easier to describe by category than by company, because the public source set does not name direct startup rivals. On one side are incumbent construction software systems that already store drawings, quality records, schedules, and site data, including the integrations Stoa itself references such as M-Files, SharePoint, Dalux, and Congrid [stoabuild.com, retrieved 2026]. Those vendors have distribution, installed workflows, and permissioning already in place, which makes them the default alternative to adding a new AI layer. On another side are internal project teams using manual search, email, and spreadsheet-based coordination as an adjacent substitute, especially in firms that are document-heavy but slow to change operating practices. Stoa's pitch, according to public coverage, is that these sources should remain the system of record while its software reads across them and answers project questions in plain language with source citations attached [Sesamers, November 2026] [stoabuild.com, retrieved 2026].
Its edge today looks real, but narrow. Public reporting supports three early advantages: a stated focus on complex construction and data-centre projects, integration into existing document systems with read-only sync, and a founding team with backgrounds spanning software, cloud, and construction-adjacent environments [Startup Researcher, November 2026] [Sesamers, November 2026] [Forbes, retrieved 2026]. That combination matters because buyers in this category tend to care about auditability and implementation friction more than about generic AI claims. Stoa's emphasis on answers grounded in customer documents, sentence-level citations, EU hosting, and a statement that customer data is not used to train foundation models is well aligned with that buying logic [stoabuild.com, retrieved 2026]. The durability question is less settled. Integration patterns and compliance messaging can travel quickly across software markets, so this edge is durable only if Stoa compounds it into deeper workflow adoption or privileged project data access before larger platforms ship similar features.
The exposure is equally visible. SharePoint and M-Files are not construction-native challengers, but they already sit near the document base; Dalux and Congrid are more category-specific and may have a shorter path to embedding AI directly into existing site workflows if customers prefer a single vendor relationship [stoabuild.com, retrieved 2026]. Stoa's public case also leans heavily on interoperability rather than system replacement, which lowers adoption friction but can weaken control over the customer relationship if the incumbent platform keeps the primary seat. The company's reported footprint of more than 25 Nordic sites by November 2026 is encouraging, but it remains early relative to the distribution advantage of established software vendors with broader installed bases [Sesamers, November 2026].
The most plausible 18-month scenario is that the market bifurcates between incumbent suites adding AI search and summarization, and specialist overlays trying to win where multi-system complexity is highest. If source-grounded, cross-system answers become the main procurement criterion, Stoa is the most obvious winner if it can show that its read-only architecture and citations reduce buyer risk without forcing workflow migration [stoabuild.com, retrieved 2026] [Terms of Service - Stoa, retrieved 2026]. The most obvious loser if that shift does not happen is the manual coordination layer inside project teams, where information retrieval still depends on individuals knowing where documents live. A second scenario is less favorable for Stoa: if category incumbents such as Dalux or Congrid move quickly enough to make AI a bundled feature rather than a separate purchase decision, the specialist overlay position becomes more perishable because the channel is not owned by the overlay vendor [stoabuild.com, retrieved 2026].
Unconfirmed -- This section relies on public product descriptions and integration references from the company website, plus limited independent reporting from Sesamers and Startup Researcher. No public source set here names direct startup competitors or verifies relative win rates.
Opportunity
Upside case
PUBLIC The prize here is not a point tool for site admins, but a credible shot at becoming the operating layer through which high-complexity construction teams interrogate project truth across drawings, schedules, contracts, quality records, and safety reports [Sesamers, November 2026] [stoabuild.com, retrieved 2026].
The headline opportunity is straightforward: if Stoa becomes the system construction teams open when they need a reliable answer about what is happening on a live project, it can move from software vendor to workflow gatekeeper. That outcome is still early, but it is not purely aspirational. Public reporting says the product is already operating on more than 25 Nordic sites as of November 2026, with a particular emphasis on data-center projects, a segment where document volume, coordination cost, and error sensitivity are unusually high [Sesamers, November 2026] [Startup Researcher, November 2026]. The product pitch also maps to a real behavioral wedge, namely asking plain-language questions against existing project records while preserving source citations and underlying permissions, rather than forcing teams into a new system of record on day one [stoabuild.com, retrieved 2026] [Terms of Service - Stoa, retrieved 2026].
That matters because construction software has often struggled when deployment requires ripping out incumbent tools or changing field workflows too aggressively. Stoa's public materials describe a read-only sync model with systems such as M-Files, SharePoint, Dalux, and Congrid, and say the service does not modify documents in those source systems [stoabuild.com, retrieved 2026] [Terms of Service - Stoa, retrieved 2026]. If that integration model works in practice, the company can position itself as the intelligence layer across fragmented project data rather than another repository competing for primary ownership.
The upside is amplified by the founding profile and investor mix. Public reporting ties the team to Supercell, AWS, SRV, and Peikko Group, and names PlanGrid co-founders Tracy Young and Ralph Gootee among angel backers alongside Lifeline Ventures [Sesamers, November 2026] [Startup Researcher, November 2026]. That does not prove execution, but it does improve the odds that Stoa can pair product velocity with category-specific pattern recognition in construction software.
| Scenario | What happens | Catalyst | Why it's plausible |
|---|---|---|---|
| Nordic control tower | Stoa becomes the default AI query and coordination layer for large Nordic construction and data-center projects, expanding from site-level deployments into portfolio-wide usage across major contractors and owners. | A flagship contractor or developer standardizes Stoa across multiple projects after early site success [Sesamers, November 2026]. | The product is already reported on more than 25 Nordic sites, and the company is explicitly focused on data-center and complex construction workflows [Sesamers, November 2026] [Startup Researcher, November 2026]. |
| Embedded intelligence layer | Stoa wins by sitting on top of incumbent document and project systems, becoming the trusted answer engine rather than replacing the stack beneath it. | Successful integrations with tools such as SharePoint, M-Files, Dalux, and Congrid reduce deployment friction and pull Stoa into daily decision-making [stoabuild.com, retrieved 2026]. | The public product architecture is designed around read-only sync, source-cited answers, and preservation of existing permissions, which is a more adoptable posture than demanding full workflow migration [stoabuild.com, retrieved 2026] [Terms of Service - Stoa, retrieved 2026]. |
| Compliance-grade project AI | Stoa turns trust and governance into a commercial wedge for European enterprise construction, particularly where teams are cautious about model training, hosting location, and answer traceability. | Enterprise buyers make EU hosting, source-grounded outputs, and customer-data isolation mandatory in procurement [Security | Stoa, retrieved 2026]. |
The compounding dynamic, if it appears, is less about a classic consumer network effect and more about workflow entrenchment. Each new site adds another instance where teams learn they can ask Stoa for the current answer instead of hunting through drawings, specs, contracts, and site records manually [stoabuild.com, retrieved 2026]. If the product consistently returns grounded answers with document-level citations, user trust should improve, and trust is the scarce input in construction decision software [Pomegra, November 2026] [Security | Stoa, retrieved 2026].
There is also a modest but meaningful distribution flywheel in the current design. Because Stoa is described as reading from existing systems without modifying them, an initial deployment can start as a narrow search-and-answer layer, then widen into broader operational use as more project teams rely on it for coordination, issue detection, and reporting [Terms of Service - Stoa, retrieved 2026] [Sesamers, November 2026]. The company is already making claims around time saved and early mistake identification, though those are still company-side or lightly corroborated signals and should be treated cautiously [Pomegra, November 2026] [stoabuild.com, retrieved 2026].
Sizing the win is necessarily scenario-based because there is no public revenue base, valuation, or independent market-sizing dataset in the source set. A useful comparable on product shape is PlanGrid, whose co-founders are among Stoa's investors and which Autodesk acquired for $875 million in 2018 [Autodesk, November 2018] (scenario, not a forecast). If Stoa were to become the AI control layer for complex European construction projects rather than only a document assistant, the upside case could plausibly exceed a historical point-solution outcome like that, especially if it captures portfolio-level usage across owners, contractors, and specialty teams. A more conservative translation is that a company with real enterprise penetration across high-value projects, sticky integrations, and trusted AI workflows could support an outcome in the high hundreds of millions of dollars, and potentially above $1 billion, if one of the platform scenarios above materializes (scenario, not a forecast) [Autodesk, November 2018] [Sesamers, November 2026] [Startup Researcher, November 2026].
Single-source, plausible -- Core opportunity claims rely on one to two public sources, with funding and deployment partially corroborated by Startup Researcher, Sesamers, and company materials; product governance details are primarily company-sourced.
Sources
From the public record
[Sesamers, November 2026] Stoa raises €3.5M pre-seed to bring AI to construction sites | https://www.sesamers.com/funding/stoa-raises-3-5m-pre-seed-ai-construction-sites/
[Startup Researcher, November 2026] Stoa Raises €3.5 Million Pre-Seed for Construction AI | https://www.startupresearcher.com/news/stoa-raises-3-5-million-pre-seed-for-construction-ai
[Today’s Startup News, November 2026] Stoa Raises €3.5M to Bring AI to Construction Sites and Data Centres | https://www.todaysstartupnews.com/startups/stoa-raises-3-5m-pre-seed-lifeline-ventures-ai-construction-data-centres
[Finsmes, November 2026] Stoa raises €3.5 Million, backed by the founders of Supercell, Wolt and PlanGrid | https://stoabuild.com/news/stoa-raises-3-5m-pre-seed
[LinkedIn, retrieved 2026] Aleksi Taipale - Hyperco | LinkedIn | https://www.linkedin.com/in/aleksitaipale/
[stoabuild.com, retrieved 2026] Stoa | AI for construction projects, answered from your own documents | https://stoabuild.com/
[Pomegra, November 2026] Stoa raises €3.5m pre-seed to bring AI to construction sites | https://pomegra.io/news/stoa-raises-3-5m-pre-seed-to-bring-ai-to-construction-sites
[CASRAI, 2026] Stoa company profile | https://casrai.org/startups/stoa
[The AI Landscape, November 2026] Stoa Raises €3.5M to Catch Construction Mistakes Early | https://theailandscape.com/news/stoa-raises-3-5m-construction-ai/
[Terms of Service - Stoa, retrieved 2026] Terms of Service - Stoa | https://stoabuild.com/terms
[Security | Stoa, retrieved 2026] Security | Stoa | https://stoabuild.com/security
[Forbes, retrieved 2026] Supercell AI Innovation Lab | https://www.forbes.com/profile/supercell-ai-innovation-lab/
Articles about Stoa
- Stoa's AI Reads the Blueprints on More Than 25 Nordic Construction Sites — The Helsinki startup, backed by Lifeline Ventures and PlanGrid's founders, is targeting the document-heavy workflows of major data-center projects.