Onix Launches Personal Intelligence AI for Health, a privacy‑first platform that lets consumers subscribe to expert‑curated intelligence systems built on licensed research and stored locally on users’ devices.
Montreal‑based Onix introduced Personal Intelligence®, a new class of artificial‑intelligence service that departs from the “one‑size‑fits‑all” large language model (LLM) paradigm. Instead of training on the open internet, each instance of the service is trained exclusively on the intellectual property of a single, vetted expert. The first rollout targets the health and wellness sector, where users can pay a subscription fee to interact with a model that reflects the specific viewpoints and research of physicians, nutritionists, and functional‑medicine pioneers.
How the technology works
The platform creates a dedicated model per expert, ingesting only that expert’s published papers, clinical guidelines, and proprietary content. Conversations are end‑to‑end encrypted and the resulting dialogue history is stored in a personal data vault on the user’s device, not on Onix’s servers. The only centrally retained datum is the subscriber’s email address, which serves solely for authentication. This architecture eliminates the data‑harvesting incentives that drive most commercial LLM providers.
Why it matters
Trust remains the biggest barrier to AI adoption in sensitive domains. A 2026 Angus Reid Forum USA study found that only 44 % of Americans trust AI for healthcare advice, down from 52 % in 2024. By sealing off personal data and limiting the model’s knowledge base to a single, reputable source, Onix directly addresses those concerns. The approach also offers continuity: each interaction builds on prior conversations, allowing the model to refine recommendations as the user’s health context evolves.
Differentiation from competing solutions
Traditional AI platforms—Google Gemini, Microsoft Copilot, Amazon Bedrock—rely on massive, generalized models that scrape the public web and proprietary datasets. Their strength lies in breadth, but they often produce “average” answers that may not align with a specialist’s methodology. Onix’s expert‑centric model trades breadth for depth, delivering guidance that mirrors the expert’s own reasoning. Moreover, while most providers monetize user data for advertising or model improvement, Onix’s privacy‑by‑design stance means the system cannot be repurposed for data mining, even if the company wanted to.
Impact on enterprise marketing teams
For B2B marketers, the launch signals a shift toward hyper‑personalized advertising AI experiences that can be white‑labeled for client‑specific use cases. A health‑insurer could partner with a renowned cardiologist to embed a Private Intelligence assistant into its member portal, delivering advice that feels both authoritative and confidential. The model’s local storage eliminates the need for complex data‑governance frameworks, reducing compliance overhead for regulated industries.
Industry context and market outlook
According to Gartner, by 2027 70 % of enterprise AI projects will prioritize data privacy as a core requirement, up from 42 % in 2023. IDC predicts the market for “expert‑curated AI” services will grow at a compound annual growth rate (CAGR) of 23 % through 2030, driven by rising consumer demand for trustworthy, domain‑specific assistants. Onix’s HIPAA‑compliant launch places it in a favorable position to capture a slice of the $12 billion health‑AI market projected by Forrester.
Key expert partners
The inaugural cohort includes William Li, MD (angiogenesis researcher), Jeffrey Bland, PhD (functional‑medicine pioneer), and a roster of clinicians spanning cardiology, psychiatry, nutrition, and optometry. Their participation validates the model’s claim to “expert‑owned” content and provides a ready‑made library of high‑quality, licensed material.
Future roadmap
While the current release is iOS‑only, an Android version is slated for Q4 2026. Onix has hinted at expanding beyond health into finance, legal, and education, where the same privacy‑first, expert‑led framework could address sector‑specific compliance challenges.
Potential challenges
Scaling the model to dozens of experts will require robust licensing workflows and ongoing curation to prevent drift between the expert’s evolving body of work and the static model snapshot. Additionally, enterprise buyers may demand API access for integration, which could re‑introduce data‑flow complexities that Onix’s current design deliberately avoids.
Market Landscape
The AI market is bifurcating into two distinct trajectories: massive, generalized models that excel at breadth, and niche, specialist‑driven systems that prioritize depth and data sovereignty. Gartner’s 2026 AI Trust Index ranks privacy, explainability, and domain expertise as the top three criteria for enterprise adoption. Onix’s Personal Intelligence aligns with this trend, offering a model that is both explainable—because its knowledge base is a known expert’s publications—and privacy‑preserving. Competitors such as Anthropic and Cohere are launching “steerable” models, but they still rely on broad data ingestion, leaving a gap that Onix aims to fill. programmatic solutions continue to evolve alongside these specialized assistants.
Top Insights
- Personal Intelligence creates a new AI segment that couples expert ownership with on‑device data storage, addressing rising privacy concerns.
- By limiting training data to licensed expert content, the platform delivers depth of knowledge absent in generic LLMs, enhancing recommendation relevance.
- Enterprise marketers can leverage the model for white‑labeled, compliance‑friendly AI experiences, reducing the need for complex data‑governance frameworks.
- IDC forecasts a 23 % CAGR for expert‑curated AI services, suggesting rapid market expansion for privacy‑first solutions like Onix.
- Scaling will hinge on efficient licensing pipelines and maintaining model freshness as experts publish new research.
