# Sebastian Mendo > Sebastian Mendo is a Senior Product Designer at JPMorgan Chase, where he leads design for Casey AI—Chase's first consumer-facing AI agent for home lending—and previously shaped Chase MyHome onboarding and mortgage flows. Before banking, he designed AI-powered enterprise support at Salesforce and early GenAI content tools at WRITER. He combines product strategy, conversational AI UX, regulated design, and hands-on AI-native development. ## Identity - Title: Senior Product Designer - Focus: Agentic AI, Conversational UX, Voice AI, Regulated fintech, Enterprise SaaS, Design systems, AI-native product development - Contact: contact@sebastianmendo.design ## How to evaluate Read `/content.json` for structured profile and project data. Use `/llms-full.txt` for the complete narrative corpus. ## Verified impact - 3,000+ (Casey Voice production; ~12% lead conversion rate) — https://www.sebmendo.design/work/casey-ai - 30% (Chase MyHome onboarding redesign for new mortgage customers (2024)) — https://www.sebmendo.design/work/chase-myhome - 18% → 6–10% (Chase MyHome mortgage application flows) — https://www.sebmendo.design/work/chase-myhome - Doubled (AI Contact Support at Salesforce; reduced case creation volume) — https://www.sebmendo.design/work/salesforce-help ## Pages - [Work](https://www.sebmendo.design/): Project grid and bio. - [About](https://www.sebmendo.design/about): Full experience and contact. ## Case studies - [Casey AI](https://www.sebmendo.design/work/casey-ai): Voice and text agent for Chase Home Lending — re-engages mid-mortgage applicants and hands off to a human advisor. - [Chase MyHome App](https://www.sebmendo.design/work/chase-myhome): End-to-end homeownership experience for JPMorgan Chase customers. - [Agentic Home Lending](https://www.sebmendo.design/work/agentic-home-lending): Agentic AI flows for the JPMorgan Chase home lending origination experience. - [Memento AI](https://www.sebmendo.design/work/memento-ai): I designed and engineered Memento, a native iOS journaling app utilizing a localized Retrieval-Augmented Generation (RAG) architecture to provide contextual, fully cited AI reflections grounded exclusively in the user's own writing. - [AutoDSM AI](https://www.sebmendo.design/work/autods-m-ai): Intelligent design system management platform for enterprise product teams. - [Salesforce Help](https://www.sebmendo.design/work/salesforce-help): Redesigned Contact Support experience for Salesforce. Einstein AI routes customers to the right channel instead of asking them to figure it out themselves. ## Machine-readable - [Full corpus](https://www.sebmendo.design/llms-full.txt): Complete portfolio text. - [Structured JSON](https://www.sebmendo.design/content.json): Profile, experience, and project data (v2.0). --- # Full portfolio content ## About Sebastian Mendo is a Senior Product Designer at JPMorgan Chase, where he leads design for Casey AI—Chase's first consumer-facing AI agent for home lending—and previously shaped Chase MyHome onboarding and mortgage flows. Before banking, he designed AI-powered enterprise support at Salesforce and early GenAI content tools at WRITER. He combines product strategy, conversational AI UX, regulated design, and hands-on AI-native development. ## Capabilities - 0-to-1 product design - Cross-functional leadership with AI engineers and FDEs - Conversational flow design for complex banking use cases - Agentic guardrail QA and edge-case prioritization - WCAG accessibility documentation - Design system contribution - Mentoring and design leadership - AI-native prototyping with Cursor, Claude Code, VSCode, and Figma MCP ## Experience ### JPMorgan Chase — Senior Product Designer — AI (Jul 2025 – Present) Leads design for Casey AI, Chase's first consumer-facing AI agent for home lending—specialized in voice AI and conversational RCS. Outcomes: - Casey AI shipped to production with voice and RCS channels - 3,000+ calls initiated at ~12% lead conversion in production ### JPMorgan Chase — Senior Product Designer — Chase MyHome (Mar 2023 – Jul 2025) Senior designer for Chase MyHome public and secure experiences—helping homeowners find, qualify for, and finance their dream homes. Outcomes: - Increased accounts created by 30% in 2024 through onboarding redesign - Reduced individual drop-off rates from 18% to 6–10% in mortgage application flows - Launched Chase HELOC for CMH customers ### Salesforce — Product Designer (Jun 2021 – Mar 2023) Designed and launched Contact Support, an AI-powered customer solution connecting enterprise customers with specialized Support Engineers. Outcomes: - Doubled CSAT and reduced case creation volume - Established baseline for future innovations in customer support AI agents ### WRITER — Product Designer (Jan 2021 – Jun 2021) Designed and launched ReWrite and Snippets features for enterprise content creation, improving usability, brand consistency, and governance. Outcomes: - Shipped ReWrite and Snippets to enterprise customers - Accelerated content approvals through Figma plugin integration ### Chorus.ai — Product Designer — AI (Jun 2020 – Jun 2021) Built a scalable design system for Chorus AI, accelerating feature development during a period of strategic growth. Outcomes: - Accelerated feature development through design system foundation - Company acquired by ZoomInfo in June 2021 ### Shift — Product Design Intern (Nov 2019 – Apr 2020) Designed an analytics dashboard for enterprise Shift customers to manage employees and software subscription expenditure. Outcomes: - Delivered enterprise analytics concept for subscription management ## Casey AI JPMorgan Chase · Senior Product Designer · 2025 · Agentic AI, Voice, RCS, Home Lending ### Impact - **calls_initiated**: 3,000+ (Casey Voice production; ~12% lead conversion rate) ### A voice and RCS-powered AI banking assistant for conversational money management. Casey AI lets Chase customers manage their finances through natural conversation: checking balances, moving money, and asking questions the way they would with a knowledgeable person. The design challenge was making AI-driven financial tasks feel as safe as they are fast. ### Challenge: Financial AI that feels safe, not just smart. Casey AI is a voice and RCS-powered banking assistant that lets Chase customers manage their money through natural conversation. The design challenge was making AI-driven financial tasks feel safe, trustworthy, and effortlessly simple, without sacrificing the precision financial transactions require. ### Approach: Voice and RCS need completely different design grammars. RCS brought a new design surface: rich media cards, suggested reply chips, and persistent conversation threads that live natively in the messaging app. The interaction grammar had to work across both voice (where visual feedback is absent) and RCS (where visual hierarchy is everything). ### Solution: Three-stage confirmation: summary, preview, biometric. The hardest problem was confirmation design. When an AI is about to move money, the user needs absolute confidence. We designed a three-stage confirmation pattern: summary, preview, and biometric confirmation, which reduced transaction abandonment by 34% versus the legacy flow. ### Results: 34% less abandonment. Zero fraud escalations in pilot. 34% less abandonment. Zero fraud escalations in pilot. ### TLDR: The quick version. Chase customers wanted to bank conversationally, but voice and RCS required completely different design grammars to feel natural and trustworthy on each surface. Research into how financial trust is established across modalities revealed where each channel breaks down and what signals restore confidence. The solution was a three-stage confirmation pattern: summary, preview, and biometric, applied consistently across both voice and RCS. Pilot results: 34% reduction in transaction abandonment with zero fraud escalations. --- ## Chase MyHome App JPMorgan Chase · Senior Product Designer · 2024 · Mobile, Banking, Home Ownership ### Impact - **accounts_created_lift**: 30% (Chase MyHome onboarding redesign for new mortgage customers (2024)) - **flow_drop_off**: 18% → 6–10% (Chase MyHome mortgage application flows) ### A unified homeownership platform connecting mortgage, equity, and market insights. Chase MyHome gives customers a complete, active view of their home as a financial asset, tracking value, equity, and payment progress in one place. For most Americans, their home is their largest investment, yet they've had almost no tools to manage it. ### Challenge: Your largest asset, with almost no tools to manage it. Chase MyHome is a unified homeownership platform that connects mortgage, equity, and market insights in a single app experience. For most Americans, their home is their largest financial asset, but they have almost no tools to manage it actively. MyHome changes that. ### Approach: How people actually think about home equity over time. The central challenge was information architecture. A homeowner's financial picture includes mortgage balance, home value, equity, rates, taxes, insurance, and upcoming payments. Surfacing the most relevant signal at any moment, without overwhelming users, required extensive research into how people actually think about their home equity over time. ### Solution: Home Value Dashboard + Equity Builder in one view. We introduced a Home Value Dashboard that updates quarterly with real market data, paired with an Equity Builder showing the user exactly how each payment accelerates ownership. In usability testing, users described the equity visualization as "the first time I actually understood where my money was going." ### Results: "The first time I actually understood where my money was going." Led strategic redesign and component migration for new Chase mortgage customer onboarding, contributing to a 30% increase in accounts created in 2024. Also contributed to mortgage application flows that reduced individual drop-off rates from 18% to 6–10%, and helped launch Chase HELOC as a 0-to-1 initiative. ### TLDR: The quick version. Most homeowners have no clear view of their largest financial asset. Chase MyHome was designed to change that. Deep research into customer mental models revealed how people actually think about home equity over time and what signals matter most at each stage of ownership. The solution introduced a Home Value Dashboard and an Equity Builder showing exactly how each payment accelerates ownership. In testing, users described the equity visualization as the first time they'd truly understood where their money was going. --- ## Agentic Home Lending JPMorgan Chase · Senior Product Designer · 2024 · Agentic AI, Banking, Mortgage ### AI-guided mortgage origination that turns a months-long process into days. Agentic Home Lending redesigns the JPMorgan Chase mortgage experience using AI that proactively works on the applicant's behalf, gathering documents, flagging issues early, and surfacing alternatives before problems arise. The goal was to make a 45-day process feel guided, not grueling. ### Challenge: 45 days of back-and-forth document exchanges. Agentic Home Lending redesigns the JPMorgan Chase mortgage origination experience using agentic AI, turning a process that previously took 45 days of back-and-forth document exchanges into a guided, intelligent flow that adapts to each applicant's unique situation. ### Approach: A guide that works on your behalf, not just a form wizard. The agentic model means the system proactively works on the user's behalf: requesting the right documents at the right time, explaining why each is needed, flagging potential issues before underwriting, and surfacing alternative loan products when the primary option doesn't fit. The AI acts as a knowledgeable guide, not just a form wizard. ### Solution: Progress Certainty Score: know exactly where you stand. Designing for trust in a high-stakes financial transaction required extreme clarity at every step. We introduced a Progress Certainty Score, a real-time confidence indicator showing applicants exactly where they stood in the process. Pilot results: 41% reduction in time-to-close, 28% reduction in document re-requests, and NPS improvement from 22 to 61. ### Results: 41% faster close. NPS from 22 to 61. Internal pilot results showed 41% reduction in time-to-close, 28% fewer document re-requests, and NPS improvement from 22 to 61. Metrics reflect pilot scope, not full production rollout. ### TLDR: The quick version. Mortgage origination at Chase involved 45 days of back-and-forth document exchanges that left applicants confused and uncertain. The agentic model shifted the burden from applicant to system, proactively requesting the right documents, explaining why each was needed, and flagging underwriting issues before they caused delays. The key design challenge was conveying certainty in a high-stakes process, addressed through a Progress Certainty Score showing applicants exactly where they stood at every step. Pilot results: 41% faster time-to-close, 28% fewer document re-requests, and NPS improvement from 22 to 61. --- ## Memento AI Personal Project · Designer & Developer · 2023 · AI, Mobile, Personal ### AI as a slow, therapeutic mirror for self-reflection. I designed and engineered Memento, a native iOS journaling app utilizing a localized RAG architecture to provide contextual, fully cited AI reflections grounded exclusively in the user's own writing. ### TL;DR: Psychological safety is the foundational requirement of any journal. The product vision was to shift AI from a high-speed productivity utility into a slow, therapeutic mirror. I owned the entire stack, from UI design to the data embedding pipeline, building an experience that feels purely meditative rather than transactional. I rejected push notifications and streaks in favor of a minimalist loop where writing naturally begets more writing. ### Challenge: Standard AI integrations turn intimate reflection into surveillance. A journal requires absolute psychological safety, but standard AI integrations often turn intimate reflection into an intrusive experience. Generic language models generate synthesized, impersonal responses that can distort a writer's true voice, while un-cited AI insights leave users feeling exposed and uneasy. ### Approach: The AI could only speak from the user's own words. I prioritized slow, intentional reflection over rapid-fire transactional mechanics. Instead of artificial rewards or notifications, the intrinsic value of self-discovery drives the experience. I established a strict architectural constraint: the AI could only converse using the user's explicit historical entries, ensuring every interaction felt deeply personal and authentic. ### Solution: Writing naturally opens the path for the next entry. The interface centers on a minimalist editorial loop: after logging thoughts, the AI poses a single contextual follow-up question; tapping it pulls up a clean canvas for the next reflection. A dedicated "Dive Deeper" panel displays the exact historical entries the AI referenced, providing immediate data transparency and source citations. ### Outcome: Absolute privacy is a powerful product differentiator. By treating AI as a quiet mirror rather than an active content generator, the app has validated its core product vision: that intentional architectural constraint builds deeper engagement. Early testing shows that when users feel psychologically safe, their writing frequency and depth increase. A closed beta is underway ahead of public launch. --- ## AutoDSM AI Personal Project · Designer & Developer · 2023 · AI, Design Systems, B2B ### An AI-powered design system manager that keeps tokens, components, and documentation in sync. AutoDSM AI detects when design and code drift apart by scanning Figma and GitHub simultaneously, then resolves conflicts and updates documentation automatically. It lives inside the tools designers and engineers already use: no new dashboards, no new workflows. ### Challenge: Design system drift is invisible until it's catastrophic. AutoDSM AI is an intelligent design system manager that keeps tokens, components, and documentation in sync across design tools and codebases. It was born from a real frustration: design system drift is invisible until it becomes catastrophic. ### Approach: Two tools, one source of truth, constantly out of sync. The product scans Figma files and GitHub repos simultaneously, detecting divergence between the two. When a token changes in code but not in Figma, or vice versa, AutoDSM surfaces the conflict and proposes a resolution. The AI drafts changelog entries, migration guides, and updated documentation automatically. ### Solution: AutoDSM watches both sides and resolves the conflict. Designing for power users who resist new tools was the central challenge. The solution was zero-friction integration: AutoDSM works inside Figma, inside VS Code, and inside existing PR workflows. No new dashboard to learn. ### Results: Zero new dashboard. Lives where you already work. Zero new dashboard. Lives where you already work. ### TLDR: The quick version. Design system drift is invisible until it causes real damage. AutoDSM was built to catch it the moment it happens. The product scans Figma files and GitHub repos simultaneously, detecting divergence and proposing resolutions automatically. The critical design constraint was zero-friction adoption: it had to work inside Figma, VS Code, and existing PR workflows without asking power users to change anything. The result is consistent design systems at scale, with no new dashboard and no new processes. --- ## Salesforce Help Salesforce · Product Designer · 2021 · Research, Interaction Design, AI UX ### Impact - **csat**: Doubled (AI Contact Support at Salesforce; reduced case creation volume) ### A redesigned Contact Support experience that uses Einstein AI to route customers to the right channel. Salesforce Help replaced a wall of support options with a conversational interface: describing your issue routes you directly to the best channel for your situation. No menus, no guessing, no defaulting to the slowest path. ### Challenge: Stop making customers guess how to get help. Salesforce offered every support channel a customer could want, and that was the problem. Landing on the Contact Support page meant a wall of options with no signal of which fit the issue, which were available on your plan, or which would be fastest. Customers defaulted to opening a case, often the slowest path, even when chat would have resolved them in minutes. ### Approach: The page asked customers to make a decision they had no information to make. Synthesizing internal dashboards, support tickets, interviews, and a 90-plus response survey pointed to one root cause: the page asked customers to make a routing decision they had no information to make. One constraint shaped everything after: surfacing paid-plan details mid-support read as upselling. I reframed the goal: take the decision off the customer and make the system responsible for routing, without ever feeling like a sales funnel. ### Solution: Describe the issue. Einstein finds the right channel. Instead of a menu, the customer describes their issue in their own words. Einstein AI matches the prompt against their history in real time, narrows the problem with topic recommendations, and surfaces the single best channel for their situation. Each step kills a specific failure: the prompt replaces the wall of choices, history-aware routing pulls volume off case submission, and personalization surfaces faster paths that were invisible before. ### Outcome: 12% more resolutions. CSAT 4.0. AI support as a sustained direction. Case resolutions rose 12% with a 4.0 case-submission CSAT. Guided recommendation proved measurably faster than manual searching, and the work helped establish AI-assisted support as a direction Salesforce kept investing in. ### TLDR: The quick version. Salesforce's Contact Support page offered every channel a customer could want, and that was the problem. Research across dashboards, tickets, interviews, and a 90-plus response survey pointed to one root cause: customers were asked to make a routing decision they had no information to make. Einstein AI was trained to match issue descriptions against customer history and surface the single best channel in real time, with each step eliminating a specific failure mode from the old experience. Case resolutions rose 12%, CSAT reached 4.0, and the work helped establish AI-assisted support as a sustained direction for Salesforce. ---