Applied AI for regulated and responsible industries.
Where we point AI, and how, matters as much as whether we use it at all.
We practice augmentation, not replacement. AI absorbs the scale and complexity no team can carry, so human judgment can be spent where it matters. The value comes from solving specific bottlenecks, not from grand ambitions. And the people a system affects get a voice in its design, early.
Read the case for augmentation
Regulated and responsible AI.
The durable advantage in AI is not the model. It is the discipline around it. Everything we build is designed to pass the scrutiny of an auditor, a regulator, and the people who use it every day.
Knowledge you can question
Retrieval systems that answer plain-language questions against controlled, auditable knowledge bases without exposing sensitive data. Proven in live client environments.
Governance, risk, and compliance
With CGI's cybersecurity practice, which already delivers AI GRC, we help clients meet the expectations of boards and regulators before AI reaches production.
Delivery that holds up
We prove capabilities on CGI's own operations first, then carry them to clients through phased rollouts, working within approved tooling and governance from day one.
What we offer.
Three categories of service: technical builds, advisory work, and the enablement that makes adoption stick.
Technical
The AI systems we design, build, and supportAssistants that answer questions over policies, procedures, and institutional knowledge, with sources.
Systems that draft, assemble, and standardize proposal and business-development material.
Workforce intelligence that predicts availability and matches people to the right work.
Pipelines that turn raw source material into branded, formatted, review-ready deliverables.
Tooling that decodes large operational reports and document sets into answers teams can act on.
Cost-conscious cloud architecture and re-architecture for AI and data platforms.
Advisory
Strategy, governance, and readinessGovernance, risk, and compliance for AI, delivered with CGI's cybersecurity practice.
Structured assessment of data, appetite, and risk posture that turns interest into a sequenced plan.
Roadmaps and compliance tracks for organizations whose data must be in order before AI can be.
Facilitated sessions that align leadership on priorities and the use cases worth funding.
Enablement
Building capability and confidenceTrain-the-trainer programs and hands-on sessions, from AI fundamentals to machine learning.
Speaking at executive summits, industry conferences, and community events.
Hackathon mentorship, student team leadership, and a standing presence in the regional AI community.
We work where the rules are real.
Our engagements span banking, insurance, healthcare, exports, and other regulated and responsible sectors. Different industries, one constant: the systems we build have to answer to someone.
Banking and credit unions
Member data, audit trails, and regulator expectations at the core of every build.
Insurance
Underwriting knowledge, policy documents, and readiness for AI under compliance.
Healthcare
Privacy-first delivery inside approved clinical tooling and governance frameworks.
Exports and manufacturing
AI adoption for executive teams in inspected, certified, and traceable industries.
Government and public sector
Approved tooling, procurement discipline, and public accountability by design.
Education
Applied prototypes and institutional data work with academic partners.
Discovery first. The findings choose the path.
We do not arrive with a solution looking for a problem. Every engagement starts with discovery, and what we find determines what happens next.
A facilitated look at your priorities, data, risk posture, and the standards you answer to. Short, structured, and honest: if AI is not the right answer yet, discovery is where we say so.
What we find leads to one of three paths
Advise
When the foundation needs workIf the data is not ready or governance questions are unanswered, we start there: data strategy, AI GRC, and a sequenced plan that earns the right to build.
Build
When the use case is soundIf the data is solid and a use case scores well on value and risk, we build a proof of value and carry it through to a system in production.
Enable
When the gap is confidenceIf the missing piece is skills or trust, we start with training, workshops, and safe adoption, so your people grow with the technology instead of around it.
Most engagements combine paths. Discovery decides the order.
Two data scientists, one vision.
AIx is a women-led lab pairing production delivery with active research and a standing presence in Atlantic Canada's AI community.
Rose Boudreau
Rose is a Data Scientist and Co-Lead of AIx with CGI's Atlantic Business Unit and the architect of CGI Arc, an enterprise retrieval-augmented generation platform deployed for banking and federal clients. She holds a Data Analytics diploma from Nova Scotia Community College and is pursuing a Master of Science in Artificial Intelligence from the University of Colorado Boulder, researching adversarial interpretability of large language models and socially situated behaviour. She is also the founder of Quinan Labs, a research validation lab that has secured over $150,000 in non-dilutive funding to validate AI deployment across industry, government, and community contexts.
Sana Kashgouli
Sana is a computer scientist specializing in scalable, data-driven AI solutions and advanced algorithm design. She holds a Master's in Computer Science from Dalhousie University with a focus on Data Structures and Algorithms, and brings deep experience across AI, large language models, information retrieval, and cloud strategy. Sana has conducted first-author research on AI-driven knowledge extraction, served as an AI Researcher on interdisciplinary projects with the National Research Council Canada, contributed to genomic data research in compression and pattern matching, and presented her work at WiMLDS, the ECS Conference, and international workshops.
If your industry answers to a regulator, we should talk.
A 90-minute executive briefing is the usual starting point. Bring the workflow that hurts and the standards you answer to. We will bring what has already worked in environments like yours.
