Product strategy / Experience architecture / Applied AI
PORTFOLIO / 2026I’m Jay Sethi. I connect strategy, customer experience and the systems underneath, turning ambiguous business problems into clear decisions, useful products and work a team can build on.
20+ years across agency & enterprise.
Selected employment and client work.
Jay Sethi · Principal Product, Experience & AI Strategist
Toronto, Canada / Remote · jay@jsethi.com · jsethi.com
Different problems.
The same connective thinking.
Find the real constraint. Make the decisions explicit. Build a better way forward.
A RONA component I designed revealed a useful contradiction: reassurance that offered none. I turned the finding into a clear display rule, implementation brief and repeatable diagnostic approach.
Experience-led diagnosis · Six diagnostic frameworks
I built an AI-assisted research harness that retains sources, applies consistent criteria and turns a competitive assessment into a repeatable intelligence capability.
11 carrier ecosystems12 synthesis dashboards
Research architecture · Information architecture · AI workflows
A local prototype connecting briefs, supportable claims, draft assembly, automated checks and human review. The goal: accepted work, not simply more generated variants.
4 human decision gatesExecutable local prototype
Product strategy · Service design · Hands-on prototyping
I built the evidence pipeline and Reader, then designed a first application for pharmaceutical marketing: preparing source-linked evidence packages for review.
292,649 records in dry replay30.7M source bindings checked
Built foundation · Proposed marketing activation workflow
A rebuildable knowledge graph connecting campaigns, brands and agencies to supporting sources and making uncertain interpretations visible rather than hiding them.
380K entities1.2M relationships
Independent R&D · Information architecture · Quality systems
From “what should we do?”
to “here’s how it works.”
Frame the opportunity, understand the customer and align people around the decisions that matter before the roadmap hardens.
Connect journeys, information, content and component behaviour so a useful idea survives the details of implementation.
Turn useful methods into AI-assisted tools and workflows, with clear source evidence, testable rules and human judgment in the right places.
Senior enough to align the room.
Hands-on enough to move the work.
I bring 20+ years across product strategy, UX, information architecture and client leadership. My work spans commerce, financial services, insurance, healthcare, pharma and enterprise software.
Today, I combine that foundation with hands-on AI-assisted development: framing opportunities, designing the experience, building testable systems and making the tradeoffs clear to the people who need to act.
My value is the connection between those disciplines, not just one deliverable at the end.
Request a tailored résuméConsulting and product development across conversion diagnostics, agency workflows, competitive intelligence and source-backed information products.
Connected commerce journeys across navigation, search, product pages, cart, checkout and fulfilment.
Digital banking direction, roadmap ownership and alignment across an approximately 30-person cross-functional organization.
Client-facing discovery, experience architecture and direction of multidisciplinary work across enterprise and life-sciences engagements.
Managed a five-person design team while remaining hands-on with enterprise software experiences.
MaRS Discovery District: UX Design Lead
General Assembly: UX & Product Design Instructor
Baywood FHO: Digital Transformation Consultant
Brightworks: Information Architect
TUI / First Choice Travel: UX Designer
The ideas underneath the projects.
The State of Design 2026 ·
A strategic field guide to thirteen design and AI-building tools: agent access, review controls, economics and portability. Compare the workflow, not just the demo.
Documentation-based review · Editorial recommendations
Read the field guideEU KIDS Act · Strategic playbook ·
A role-based playbook for product decisions, minimal-data age assurance and governance under the proposed EU KIDS Act. Prepare for change without hard-wiring a draft into your product.
Proposal-focused analysis · Implementation recommendations · Not legal advice
Ambient macros & EHRs · Product intelligence ·
A 14-vendor comparison of reusable macros, chart-data binding and ambient AI and what those differences mean for product direction, workflow design and clinician control.
Vendor landscape · Methods & comparison · Proposed product direction
Jev & SOAP notes · Field note ·
Where I would test TypeSafe’s Jev to organize a conversation, choose reusable note components and recognize when more context or a better template is needed.
Research-informed proposal · Not a tested clinical integration
The useful question is not only how quickly a model can generate variants. It is whether strategy, evidence, creative judgment and review can move together without losing the context that makes the work usable.
I’m interested in connecting those handoffs: reusable briefs, explicit claims, bounded generation, meaningful checks and human decisions. The output is an operating capability, not simply a larger pile of content.
See this thinking in the AI sprintA system can retain the page a fact came from and still misunderstand the fact. Source identity, document versions, meaning and review decisions deserve separate attention.
That distinction shapes my evidence systems: make the supporting passage inspectable, preserve uncertainty and test the interpretation, not just the plumbing. Better prompts cannot substitute for a clearer information model.
See the lesson from AIKGFor a person, useful content needs a clear purpose, a meaningful place and a next action. For a machine intermediary, the source, entity, context and relationships need to remain legible too.
I treat those as connected design questions. Can the person find what matters? Can a system retrieve it? Does the context survive the handoff? My competitive research keeps those questions separate enough to answer honestly, and connected enough to inform experience strategy.
Explore the competitive harnessA confusing interface may be expressing an unclear rule, an awkward handoff or a mismatch between content and intent. Redesigning the surface is only part of the work.
The RONA warranty example is small but telling: the fix is not more reassuring copy for “None.” It is a component rule that only shows optional reassurance when there is something truthful to reassure the customer about.
See the component decisionA macro trigger is only one decision. I would also test where a statement belongs in a SOAP note, which reusable component fits, and whether it needs a supported modifier, another component or more context.
Adding a stated qualifier is not the same as filling a gap. Repeated edits may reveal that a template needs redesign, but fixing today’s note and changing the shared library should remain separate, reviewed decisions.
Jev would help propose those choices, not write undocumented findings or authorize clinical actions. I would judge the workflow by preserved meaning and less clinician correction, not by how much it automates.
The full field note develops the idea with a synthetic conversation, four kinds of component mismatch and a practical evaluation plan.
Research-informed proposal, not a tested clinical integration. The full article includes the evidence and limitations.
Read the full Jev & SOAP field noteA customer journey. A new product. A team’s next chapter.
Let’s start with the problem and make the next move clear.
Email: jay@jsethi.com · Web: https://jsethi.com · LinkedIn: https://www.linkedin.com/in/jaysethi