oction labs
Brand + investor deck
Secure AI deployment and vetted data for regulated industries
Brandon Gill, Co-Founder  ·  MJ Dewji, Co-Founder  ·  June 2026
Deployments live $25M seed / pre-series A Confidential Q3 2026 target close
brandon@octionagency.com  ·  mj@octionagency.com octionagency.com
Foundation
"We deploy AI inside regulated environments, de-identify and vet the data those environments generate, and license the resulting datasets to companies training the next generation of AI models."
Layer 1
Secure AI deployment
Layer 2
De-identification pipeline
Layer 3
Vetted data licensing
The problem
Regulated industries cannot legally use general-purpose AI
  • Municipal governments — provincial data-sovereignty laws prohibit cloud LLMs for citizen data.
  • Healthcare — HIPAA and PIPEDA require on-premise handling of patient records, diagnoses, and care history.
  • Finance and legal — SOC 2 and client confidentiality rules block SaaS AI tools from touching case or account data.
  • Public safety — NIST 800-53 mandates controlled supply chains, audit trails, and data-residency guarantees.
The gap
Every AI vendor pitches "secure." None can prove sovereign infrastructure. Regulated buyers want AI. They cannot legally buy what exists.
Our answer
AI that runs inside your environment. Your data never leaves. Decisions are traceable and auditable.
What we do
Three layers. Each one compounds the last.
Layer 1
Secure AI deployment
We deploy AI systems inside regulated environments: healthcare, government, finance, legal. The software runs on the client's own infrastructure or our nodes via secure tunnel. Data does not leave the environment. NIST 800-53, HIPAA, PIPEDA compliance built in.
Revenue: deployment and managed services fees
Layer 2
De-identification and vetting
As the system operates, data interactions are anonymized in real time. We classify every data type by what can and cannot be shared under applicable law. The result is a structured, verified, shareable dataset built from real regulated-industry activity.
Moat: no other party has legal access to this data
Layer 3
Vetted training data licensing
The curated datasets are licensed to AI companies building domain-specific models. Healthcare AI companies, public sector AI vendors, and enterprise AI developers all need training data from regulated environments. There is no compliant alternative source.
Revenue: recurring data subscriptions and one-time licenses
The niche strategy: we go deep in a small number of regulated sectors before expanding. Depth makes the data defensible.
Why now
Three converging forces create a 24-36 month window
01   Regulatory urgency
NIST 800-53 Rev.5, provincial AI procurement rules, and EU AI Act spillover into Canadian standards are creating mandatory procurement criteria that only sovereign infrastructure can satisfy.
02   Training data scarcity
The next wave of AI model development requires domain-specific, high-quality training data from regulated environments. This data is locked behind compliance walls that no cloud AI company can cross. Oction is already inside.
03   Talent arbitrage
Calgary and Edmonton AI ecosystems are significantly under-recruited relative to Toronto and Vancouver. We operate there. Competing firms have not discovered this yet.
Miss this window and incumbents (Deloitte, CGI, IBM) will fill it with overpriced, underperforming solutions.
Market opportunity
Total addressable in 5 years: $1.1B
SegmentTAMSAM (5-year)Notes
Regulated AI deployment services$8.4B$420MHealth, government, finance, legal - managed services
AI training data licensing$4.1B$380MVetted, de-identified regulated-sector datasets
Canadian municipal AI$2.1B$180MPredictable procurement, defined compliance path
Domain model fine-tuning data$1.8B$120M (Year 3)Niche-specific model weights and curated corpora
The deployment business funds operations. The data licensing business is where the margin compounds. No competitor is positioned across both.
Our identity
Infrastructure that earns trust before it earns revenue.
Oction is not a product company. We are an infrastructure company. The brand reflects that: precise, dark, and functional. One point of colour, used only where it matters.
Dark by default
We design for engineers, not pitch rooms. Precision over decoration. No gradients. No hype. No exclamation marks.
Sovereign by design
The hardware is ours. The data stays yours. The compliance is built in, not claimed. This is not a feature. It is the architecture.
One signal
Coral is the only accent in our design system. It is used once, where it matters most. That discipline is the same discipline we bring to every deployment.
The pipeline
Secure deployment is the entry point. The data asset is the business.
Year 1: deployment fees. Year 3: data licensing at margin. Year 5: the most defensible regulated-sector dataset in Canada.
Depth in a few niches beats breadth across many.
1
Secure deployment
Oction deploys AI inside the regulated environment. All processing stays within the client's infrastructure or our secured nodes. No data egress.
2
Real-time de-identification
Every interaction is anonymized as it occurs. PII, PHI, and identifying signals are stripped at the point of generation, not after the fact.
3
Shareability classification
Each data type is assessed against applicable law and client consent agreements. A binary determination is made: shareable or not shareable. No grey zone.
4
Vetted dataset licensing
Curated, verified, sector-specific datasets are licensed to AI companies. The buyer gets legally defensible training data from environments they cannot access themselves.
Competitive landscape
No direct competitor for the full stack
CompetitorTheir weaknessOur position
OpenAI / Anthropic / GoogleCannot sign data-residency agreements. PHI/PII cannot leave the facility.We operate inside the environment. They cannot.
Deloitte / CGI / Accenture$300+/hr labour, no AI expertise, no data pipelineSoftware-first, faster, and we retain the data asset
AI training data brokersPublic web scrapes, synthetic data, no regulated-sector accessReal, verified, de-identified data from live regulated deployments
Local MSPs and health IT vendorsNo AI expertise, no compliance fluency, no de-identification pipelinePurpose-built for compliance from day one
Moat: compliance fluency × inside access × de-identification pipeline = the only source of legally defensible regulated-sector training data in Canada.
Team
Who builds and operates the system
Brandon Gill
Co-Founder, CEO
Sales, partnerships, and capital strategy. Manages client relationships and government alignment.
MJ Dewji
Co-Founder
Strategic oversight, capital deployment, and long-term positioning.
Leonel Garcia
Partner, COO
Runs all operations. Oversees deployment execution, client delivery, and operational systems.
Shamir Cheema
Partner, CRO
Revenue strategy, pipeline development, and commercial growth across regulated sectors.
AI staff (active, 24/7)
Lucius - Operations intelligence Atlas - Security and forensics Avery - Research Kai - Data Oscar - Content Finn - Sales
We operate around the clock. We do not burn out. We do not leave for Google.
The bet
We are already inside the most restricted data environments in the country.
We de-identify it, vet it, and classify exactly what can leave.
Companies building the next generation of AI models need this data. There is no other legal way to get it.
The deployment business gets us inside. The de-identification pipeline builds the asset. The licensing business is the outcome.

Every niche we go deep on makes the dataset harder to replicate and more valuable to the buyers who need it.
The plan
What changes the day funding closes
First priority   Hiring
Build the team that scales the deployments
The constraint today is people, not opportunity. With funding we hire the deployment engineers, compliance specialists, and client success operators to run multiple engagements in parallel. Each niche needs dedicated staff who understand the regulatory environment, not generalists.
Second priority   Compute
Expand the node infrastructure
More compute means more concurrent deployments, faster de-identification throughput, and deeper knowledge indexing per client. Our current nodes handle the proof of concept. Scaled infrastructure handles the pipeline at commercial volume.
Third priority   Niche depth
Go deeper before going wider
Healthcare and municipal first. Depth builds a dataset that is hard to replicate. Once those two niches are established, the model extends to finance and legal.
The ask
$25M  ·  Seed / Pre-Series A

Services revenue funds operations. Capital funds the people and compute that turn the proof of concept into a company.
oction labs
Secure deployments today. The most valuable regulated-sector dataset in Canada tomorrow.
Brandon Gill  ·  brandon@octionagency.com
MJ Dewji  ·  mj@octionagency.com
Leonel Garcia  ·  leonel@octionagency.com
Oction Labs Sovereign AI deployment Vetted data licensing octionagency.com