Aakash GuptaHow to Build AI Products in FinTech ($100B Robinhood VP Lessons)
CHAPTERS
Robinhood at $100B: why this conversation matters
Aakash frames Robinhood’s recent surge in market cap and introduces Abhishek Fatehpuria, the VP of Product behind many core brokerage experiences. The episode is positioned as a rare look into Robinhood’s product-building playbook in a highly regulated domain.
- •Robinhood crosses $100B market cap and strong stock performance context
- •Introducing Abhishek Fatehpuria and the focus on product-building lessons
- •Promise of behind-the-scenes detail on how Robinhood ships products
Cortex AI assistant: solving “why did this stock move?” in the user workflow
Abhishek explains Robinhood Cortex and the first shipping use case: Stock Digest. The goal is not to add AI for novelty, but to compress the investor’s research loop directly inside moments users already experience in-app (e.g., reacting to price-move notifications).
- •Cortex starts with Stock Digest to explain stock price moves
- •Pulls from news, research reports, Robinhood trading context, and SEC filings
- •Designed to fit existing workflows rather than create a new AI destination
- •Rollout begins with stocks, with crypto planned next
AI guardrails in FinTech: trust, licensing, and “no recommendations (yet)”
The discussion turns to what makes AI product development different in finance: regulatory constraints and the need for customer trust. Abhishek outlines their approach—curate inputs, control outputs, and avoid crossing into advice until the bar can be met.
- •Upstream data curation: licensed news, research, exchange data, filings
- •Prompting/coaching to reduce hallucinations and prevent recommendations
- •Current posture: informational tool; future possibility: advisory/recommendations
- •Sequencing rationale: recommendations require portfolio context and higher safety bar
From AI demos to product roadmap: Trade Builder and problem-first AI strategy
Abhishek shares Cortex’s broader concept beyond Stock Digest, including a Trade Builder that helps turn a hypothesis into an executable trade. The unifying theme is starting from validated customer problems and designing AI around them—not building AI “because AI.”
- •March announcement included Stock Digest and Trade Builder concepts
- •Trade Builder: price targets, analysis, screening, and mapping views to trades/options
- •Customer-problem-first approach to selecting AI use cases
- •Internal adoption: Cortex also helps Robinhood teams answer market questions
Advice for FinTech PMs building AI: regulatory fluency, realism, and patience
Abhishek generalizes lessons for PMs: understand what the tech can do, understand what regulation allows, and progress incrementally. He emphasizes that trust in money-related AI is earned step-by-step for customers, legal partners, and regulators.
- •Build dual literacy: AI capabilities + regulatory constraints
- •Don’t surprise legal/compliance—arrive with informed, grounded proposals
- •Start small, compound comfort and trust over time
- •Customer trust is a gating factor in AI adoption for financial decisions
Tokenization explained: stock tokens and demand for private-company access
Aakash asks about the Cannes presentation where Robinhood announced tokenization efforts. Abhishek explains stock tokens as “stablecoins for stocks” to improve international access, and highlights strong retail demand for investing in private companies that stay private longer.
- •Stock tokens aim to make U.S. equities more accessible internationally
- •Analogy: stablecoins increased access to USD; stock tokens could do similar for stocks
- •Private stock tokenization taps into demand to invest in companies like OpenAI pre-IPO
- •Retail investors feel increasingly locked out as more value accrues in private markets
IPO Access: how Robinhood gets retail into IPO allocations
Abhishek tells the origin story of IPO Access and how the product works operationally. Robinhood joins the selling group, collects retail indications of interest during the roadshow period, shares demand with underwriters, and allocates shares based on what’s granted the night before listing.
- •Built in 2020–2021, partly inspired by enabling access to Robinhood’s own IPO
- •Mechanics: collect retail orders post-prospectus, share demand with underwriters
- •Allocation depends on how many shares underwriters grant; often oversubscribed
- •Differentiation: opening IPO access beyond high-net-worth “preferred” clients
Robinhood’s innovation DNA: value + delight, plus “Swipeys” as working backwards
They zoom out to what makes Robinhood’s product approach distinct in regulated products that are hard to reinvent. Abhishek describes a consistent bar: ship products that deliver meaningful customer value and a delightful experience, and use “swipeys” (mobile onboarding screens) to force clarity early.
- •Innovation constraint: many FinTech products are legislated (e.g., IRAs) and can’t be reinvented
- •Bar: new products should deliver both strong customer value and great UX
- •“Swipeys” exercise: define the product in 3–4 screens before building to sharpen the narrative
- •Mobile-first “working backwards”: if you can’t earn a ‘Get Started’ in one sentence, rethink
Polish, pixels, and partnering with Legal without “Frankenstein” outcomes
Aakash probes how Robinhood avoids death-by-a-thousand-departments in regulated launches. Abhishek credits deep domain expertise in legal/compliance, a culture of shared product ownership, and specific behaviors: assume good intent, sell the vision, and deeply understand the underlying rule/issue.
- •Strong legal/compliance expertise (including ex-regulators) enables better product solutions
- •Culture: legal partners are collaborators, not blockers; escalation exists but isn’t default
- •Coaching PMs: assume good intent, get cross-functional buy-in to the vision, learn the real constraint
- •Most issues are gray-area tradeoffs; empathy and domain understanding unlock solutions
Robinhood scaling and key moments: joining in 2016, mission breadth, events, and the IPO
Abhishek reflects on why he joined (talent density) and what kept him (product craft and ambition). They discuss product velocity, Robinhood’s product events as a focusing mechanism, and the emotional highs of the IPO alongside brand lows during 2021 controversies.
- •Early draw: high talent density and constant learning in a post-PMF company
- •Ongoing focus: craftsmanship, product velocity, and an expansive mission (credit card, banking, global)
- •Product events/keynotes drive organizational clarity and external attention
- •IPO week memories; 2021 brand challenges and the resilience required to recover
Org structure and planning: GM business units, goals over OKRs, and strategic metrics
The episode shifts into operating model: Robinhood moved to GM-aligned business units (brokerage, crypto, money) after the 2022 RIF, bringing many functions under shared orgs. Abhishek explains planning via big bets and keynote-driven roadmaps, plus simplified goal-setting and core strategic arcs/metrics.
- •Structure: GM model across brokerage, crypto, and money; sub-businesses like futures/international
- •Post-2022 shift improved alignment by grouping product/eng/ops/compliance under one business org
- •Planning: align on big bets + targets, then teams build plans; increasingly ‘what’s in the keynote?’
- •Metrics arcs: #1 for active traders (market share), #1 wallet share for next-gen (net deposits, Gold), build global ecosystem (international customers)
- •Goal system over OKRs to avoid semantic overhead; refresh and grade quarterly
How PMs use AI internally + PRDs vs prototypes in regulated product work
Abhishek shares how he wants PMs using AI today: early ideation, research, and reducing administrative toil, while cautioning against expecting deep product insight from AI-generated PRDs. He argues prototypes and “swipeys” are better for product reviews, though PRDs still matter in FinTech for edge cases and regulatory detail.
- •AI for PMs: ideation prompts, opportunity research, and faster repetitive writing
- •AI isn’t yet a substitute for high-signal product thinking or full PRDs
- •In reviews, prototypes/mockups outperform PRDs for judging real UX outcomes
- •FinTech reality: PRDs remain necessary for rules, regulations, and edge cases
Experimentation mastery: the referral program’s evolution and why many PMs stop too early
Abhishek recounts his early growth work building Robinhood’s referral program, emphasizing relentless iteration. The biggest unlocks included switching from cash to variable stock rewards, adding a “claim your stock” action to drive activation, and follow-up nudges that reinforced ownership.
- •Referral evolution: fixed cash → variable cash → variable stocks, with extensive testing
- •Key insight: activation improved when users had to affirmatively “claim” the stock reward
- •Post-reward engagement: notifications (price moves/news) to build ownership mentality
- •Variable reward design: users care less about expected value and more about jackpot potential + stock ownership
- •Lesson: copy patterns (Amazon/Uber) deeply—deliver the ‘unit of service,’ not just $ incentives
Career and leadership lessons: rising from intern to VP and picking the next ‘inevitable’ company
The conversation closes with Abhishek’s personal trajectory—from engineering intern to VP—and the habits he credits: embracing detail work, showing up for cross-functional partners, and building trust. He also offers a framework for students: find post-PMF inevitability, responsible/lean founders, and companies that invest in people.
- •Career growth drivers: willingness to do gritty detail work and support teams in the trenches
- •Leadership mindset: treat the broader org as ‘your team,’ not just PM reporting lines
- •Finding the next Robinhood: seek post-PMF customer love/inevitability and disciplined founders
- •Look for environments that bet on and invest in internal talent development
- •Where to connect: Twitter/LinkedIn; episode wraps with host’s subscribe/newsletter CTA
