Tyce

Verified

Track: raisedWebsite ↗𝕏 @tyceaiLinkedIn ↗

potential 70 · scoop 82

Tyce builds private AI that ingests drawings, specs, bids, and contracts to flag risks using each contractor's own history. The Toronto startup raised a $2M pre-seed after Eastern Construction reported 40% faster preconstruction tasks.

Scores run 0 to 100. Potential: how likely the firm is to become a real winner, judged across the six dimension scores and held consistent with them, whether or not anyone has heard of it. Scoop: how unlikely a well-read tech investor is to have heard of it yet. High means it has not reached the outlets that reader follows, whatever the firm's quality or age. How our top potential scores did.

ShareXLinkedIn
Anuar Yeraliyev and Lylia Djait-Paulien spent 18 months cold-calling Toronto general contractors with zero construction experience before closing a $2M pre-seed from BDC Thrive and MaRS.

Snapshot

What they build
Contractor-configured AI using historical project data to help estimators win bids and flag risks.
Who buys
Construction contractors/estimators.
Comparables
AI co-pilot for construction bids.
Sub-industry
Private AI for preconstruction estimating

Toronto-based team with 1,200+ users and $50M+ project value supported.

How we found them

Surfaced in issue
Oct 26, 2026
Announced
Oct 20, 2026 (seed funding)
Round
pre seed · $2,000,000
Investors
BDC Thrive Fund (lead), MaRS Investment Accelerator Fund
HQ
Toronto, Canada
Discovery track
raised

Originating post ↗

Dimension scores

  • Team: 70 out of 100
  • Product: 75 out of 100
  • Traction: 60 out of 100
  • Market: 80 out of 100
  • Moat: 65 out of 100
  • Capital: 75 out of 100

Analysis

Anuar Yeraliyev worked as an ML engineer at Layer6 AI and co-authored a 2022 Nature Communications paper on somatic variant detection. Lylia Djait-Paulien spent nearly ten years at IBM in enterprise deal leadership roles.

The product processes drawings, specs, bids, and contracts through private RAG tuned to each firm's history. Eastern Construction cut preconstruction task time by 40 percent, with some five-hour jobs now taking 15 minutes. Tyce reports 1,200 users and more than $50 million in supported project value after its $2M pre-seed.

North American infrastructure spending in Texas, California, and New York drives higher bid volumes. Labor shortages in estimating make the capacity gains attractive. Integrations with Procore and Autodesk ease data intake. Historical bid records create growing lock-in inside each contractor's workflow.

What they do

Tyce enables general contractors and subs to index their own historical bids, estimates, drawings, specs, and contracts into a private AI knowledge system that drafts proposals in the firm’s voice, benchmarks new estimates against past projects, flags scope gaps/risks, levels sub-bids, and supports contract review.

Why now

Post-2023 LLM/document AI maturity (stronger multi-modal RAG, agent orchestration) meets sustained North American infrastructure spending (US state-level projects in TX/CA/NY; Canadian activity) and acute labor/estimating capacity shortages in construction. Privacy concerns favor private, customer-data-only configurations over general models. Construction digitization (Procore/Autodesk integrations) provides distribution on-ramps that were weaker 18, 24 months prior.

Founding story

Anuar’s prior startup built a product with users but failed to fundraise or close sales, leading to months of self-doubt; he deliberately sought a sales-oriented co-founder after ~30+ conversations and connected with Lylia (IBM enterprise deals background). The pair spent 18 months on customer discovery, cold calls to construction/civil engineering firms, pilots, then annual contracts, pivoting into the space after realizing neither had domain experience but that bid/estimating risks represented a high-value, underserved problem in a $2T+ North American industry.

Competitive position

Incumbents: Procore and Autodesk Construction Cloud (integrate with Tyce for data ingestion/export but lack the private historical RAG layer). Closest startups: PataBid (estimating), CountBricks (AI-assisted Canadian estimating), Structo (payment application agents). Tyce occupies a distinct preconstruction document-intelligence lane focused on contractor-specific voice, risk flagging, and bid defense rather than pure takeoff or post-construction tools.

Defensibility

Primary moat is per-contractor data lock-in (historical bids/estimates/contracts become increasingly valuable and sticky within the firm’s workflows) plus domain-specific document understanding (complex relationships across drawings, specs, addenda). Private data isolation is a credible differentiator vs. general models. Switching costs rise with integration depth and style consistency. No obvious regulatory license or exclusive talent moat yet; OSS/community ownership appears minimal.

Network density

Early team is small (LinkedIn lists ~6 employees total); founders bring complementary but non-overlapping backgrounds (ML research vs. enterprise sales). No evidence of multiple hires from the same prior organization (e.g., no Layer6 or IBM cluster visible). Hiring push for founding engineers in Toronto suggests deliberate expansion beyond the two founders.

Early team signals

Lead investors: BDC Thrive Fund (lead) and MaRS Investment Accelerator Fund (Ontario-focused seed fund with 200+ investments and significant follow-on catalysis). Additional accelerator/incubator ties noted (DMZ, Formwork Labs in some profiles). Customer signal: Eastern Construction (long-standing Ontario GC) provides specific testimonial on 40% time reduction.

Traction

  • Eastern Construction testimonial (40% time cut on bid/project-management tasks; 5-hour tasks to 15 minutes).
  • 1,200+ users and $50M+ project value cited consistently in funding coverage.
  • Annual subscription model with Procore/Autodesk integrations.
  • Live production use at “industry landmark” GCs.
  • Earlier Product Hunt success for the pre-pivot product.

Opportunities

  • US/Canadian infrastructure spending waves creating bid volume surge.
  • Expansion of integrations with dominant tools (Procore, Autodesk) lowering adoption friction.
  • Labor shortages amplifying value of “add capacity” for small estimating teams.
  • Potential for data network effects as more contractors contribute historical patterns (while preserving per-firm privacy).
  • North American reindustrialization/policy tailwinds aligning with founders’ messaging.

Risks / red flags

  • Founders had no prior construction experience and pivoted after a failed prior venture (sales execution risk remains a variable despite Lylia’s background).
  • Traction heavily reliant on one named customer (Eastern Construction) with quantitative claims; broader “1,200+ users / $50M+ project value” and “landmark GCs” are directionally positive but lack additional verified logos or ARR disclosure.
  • Small founding team in a capital-intensive, relationship-driven vertical; aggressive hiring of “cracked engineers” signals execution dependency on early technical hires.
  • Crowded and heating construction-tech AI lane with well-funded players; differentiation hinges on data moat that takes time to compound.
  • Potential regulatory exposure around contract risk flagging and proposal generation (liability for missed exclusions or generated language in high-stakes bids).

Hidden bets

  • Bet on private, per-contractor data isolation (customer data never used to train models for others) combined with RAG + multi-agent orchestration and knowledge graphs built from messy PDFs/Excels.
  • Job postings highlight FastAPI/PostgreSQL/Pydantic-AI/AWS backend, React Router/TypeScript frontend, Prosemirror editors, and evaluation harnesses for agents—pointing to production-grade document intelligence rather than thin wrappers.

Founders

Lylia Djait-Paulien, CEO
Prior: IBM / exec/enterprise deal leadership roles / ~10 years
𝕏 @LyliaDP
Anuar Yeraliyev, CTO
Education: BSc Honours Physics & Computer Science, University of British Columbia (2013–2018), exchange/research at ETH Zürich (2015–2016)
Key projects: Accurate somatic variant detection using weakly supervised deep learning (Nature Communications, 2022), Towards Autonomous Mining via Intelligent Excavators (CVPR Workshop, 2019)
Prior: Layer6 AI / ML/Product Engineer / 2022–2024; Anooka Health / Technical Co-founder / 2020–2021; Entrepreneur First / Founder-in-Residence / 2020–2021; Passenger AI / ML Engineer (Computer Vision) / 2019–2020; Weir Motion Metrics / ML Engineer (Computer Vision) / 2018–2019; Genome Institute of Singapore / Bioinformatics Research Intern / 2017
𝕏 @_startuphacker

Citations