By playing, you agree to uListen's Terms and Privacy Policy and YouTube's Terms of Service.
Product SchoolToast VP of Product on Shipping AI Into a 14-Year-Old Platform and Building for the Real Economy
Summaries, chapters, transcripts, and FAQs on this page are made by uListen, not YouTube.
CHAPTERS
- 0:00 – 1:47
Toast’s growth context and the real-economy product bar
Carlos opens by highlighting Toast’s recent business performance, setting a high-stakes backdrop for product decisions. Maggie frames the core product principle: understanding customers deeply so Toast builds things they’ll actually pay for.
- •Carlos cites strong quarter results and profitability as context for the conversation
- •Maggie emphasizes that customer understanding directly drives willingness to pay
- •Carlos frames Toast’s users as “real economy” operators with different expectations than tech-forward social audiences
- 1:47 – 2:25
Meet Maggie Crowley: Olympics and early Product School ties
Carlos introduces Maggie with personal context, including her Olympic background and earlier connection to Product School. The segment establishes her credibility and history in product leadership before the Toast chapter.
- •Carlos shares Maggie placed fifth in speed skating at the 2006 Winter Olympics
- •Maggie confirms the Olympic detail and her speed skating background
- •Carlos notes Maggie previously contributed to Product School while at Drift
- •Maggie reflects on how long ago that period feels
- 2:25 – 3:26
What Toast is beyond the payment moment: hardware, software, and demand tools
Maggie explains Toast as more than a point-of-sale swipe experience. She outlines the behind-the-scenes operational stack and additional products that help restaurants run and drive traffic.
- •Toast serves as a restaurant/retail point of sale used at kiosks, counters, and handheld devices
- •Toast includes in-store hardware plus software for menus, inventory, scheduling, prep, and performance tracking
- •Toast also offers products aimed at driving traffic to locations
- •Maggie notes Toast has a consumer app as well
- 3:26 – 4:46
Building AI for “superhero” operators: outcomes over AI theater
Maggie describes restaurant operators as sophisticated and time-constrained, with little patience for performative AI. She explains Toast’s job is to absorb model/tool complexity and deliver measurable outcomes like time savings or profit impact.
- •Maggie calls operators sophisticated despite slim margins and operational complexity
- •Toast aims to translate advances from model providers into practical workflows for operators
- •Maggie sets the success bar as real outcomes: making money, saving money, or saving time
- •She contrasts useful product work with “AI theater” that distracts from results
- 4:46 – 6:42
Shipping agents into an existing SaaS platform: behavior change and trust risk
Maggie explains why adding AI to an established product differs from building AI-native software. She highlights workflow inertia, change absorption limits, and the need for high quality from day one to avoid breaking trust across a large install base.
- •Existing users already have muscle memory and established workflows, so AI introduces behavior change challenges
- •Teams must consider how much change customers can absorb and whether the AI feature fits the outcome they want
- •Maggie warns that launching slow/inaccurate tools can burn trust at scale
- •She contrasts early-adopter tolerance for broken experiences with Toast’s higher initial quality bar
- 6:42 – 7:42
Toast IQ’s early build method: design partners, WhatsApp feedback loops, and mobile-first
Maggie walks through how Toast started experimenting with an assistant: a small cohort, tight feedback channels, and a narrow initial use case. She describes a year of iteration before scaling the approach to other major features.
- •Toast began work at the end of 2023 without full certainty on the final use case
- •The team recruited a small set of design partners and worked with them in WhatsApp groups
- •They launched first on mobile and focused on a reporting use case
- •Maggie describes iterating in a “walled off” way before scaling to broader release patterns
- 7:42 – 9:05
Driving adoption by “bringing AI to the user”: entry points, blank-page friction, and re-onboarding
Maggie argues AI features won’t succeed if teams assume users will discover them. She explains the ongoing onboarding challenge: reducing blank-page friction, making entry points contextual, and repeatedly re-onboarding as capabilities evolve behind the scenes.
- •Maggie says simply adding a chat box isn’t enough; teams must actively bring AI to users
- •She highlights the “blank page problem” and the need to teach users when to use AI
- •Because improvements are often invisible, users may not notice added capability without guidance
- •Maggie notes AI products require continuous re-onboarding due to constant change
- 9:05 – 11:02
Context-aware onboarding: role, time of day, device, and tailored prompts
Maggie details how Toast designs onboarding around the realities of restaurant work. She explains tailoring prompts and entry points based on where someone is, what they’re doing, and whether they’re on mobile or desktop.
- •Maggie says solving onboarding depends on deeply understanding how customers work day-to-day
- •She emphasizes physical context in a restaurant, time of day, and the user’s role
- •Toast IQ is accessible via multiple entry points across the product, not just a single surface
- •Toast uses pre-written starter prompts and tailors them for mobile vs desktop and different moments
- 11:02 – 12:57
Where AI UX is headed: from “help me” to “do it for me,” and the adoption clock
Carlos and Maggie discuss shifting expectations toward end-to-end task delegation. Maggie describes a lag between consumer AI interaction patterns and B2B adoption, and how product teams must anticipate when those patterns become table stakes.
- •Carlos observes user expectations shifting from assistance to full task completion
- •Maggie discusses the appeal of handing off workflows end-to-end, with unknown timing in B2B
- •She says operators don’t care whether it’s AI—only whether it solves a specific business problem
- •Maggie describes a “clock” from consumer pattern emergence to enterprise expectation
- 12:57 – 14:04
Staying relevant amid rapid change: flexibility as models and patterns evolve
Maggie reflects on how AI makes product building more dynamic and uncertain compared to earlier eras. She highlights the need to remain flexible because both interaction patterns and model capabilities are moving targets.
- •Maggie says teams must plan for shifting interaction patterns and changing model capabilities
- •She frames the current moment as uniquely interesting because product direction is less predictable
- •Maggie notes building felt less exciting before AI, but now feels more dynamic
- •Carlos adds that AI breaks traditional UI constraints by enabling new actions on demand
- 14:04 – 16:08
Market landscape vs. customer obsession: competing by embedding with operators
Carlos asks how Toast thinks about a broad ecosystem of restaurant tools and competitors. Maggie argues that trying to game out winners can distract from the fundamentals: deeply understanding users through direct, physical proximity and building accordingly.
- •Carlos lists broader category players and adjacent tools across the restaurant workflow
- •Maggie says the market outcome is uncertain and easy to over-focus on competitively
- •She argues the company that wins solves the right problems best by deeply understanding its market
- •Maggie stresses embedding with customers: visiting locations, observing work, and being directly reachable
- 16:08 – 17:16
Restaurant P&L realism as a growth constraint: earning every additional dollar
Maggie explains why pricing and growth require a tight link to restaurant economics. She emphasizes that products must justify their cost within slim margins, and that discipline is reinforced by Toast’s internal financial rigor.
- •Carlos reiterates strong financial performance and asks about growth levers
- •Maggie ties growth to building what customers truly need and will pay for
- •She stresses the need to understand restaurant P&L and where additional spend can come from
- •Maggie notes Toast’s disciplined approach, including a strong finance team, to ensure value is real
- 17:16 – 18:57
Positioning shift to “grow your restaurant” and the AI marketing breakthrough
Carlos brings up Toast IQ’s positioning shift from POS management to growth. Maggie describes marketing as a major success area, including an example of reduced agency spend and improved attribution, and explains why AI is valuable when it removes work operators don’t want to do.
- •Carlos cites the shift in positioning toward helping restaurants make more money
- •Maggie says the marketing product has been highly successful for customers
- •She shares an example of reduced marketing agency spend by over 70% and tying marketing to sales attribution
- •Maggie describes AI’s sweet spot as removing unpleasant/expensive work like email marketing and website tasks
- 18:57 – 20:31
Toast IQ Grow and model/tool abstraction: doing the complexity so customers don’t have to
Maggie outlines Toast IQ Grow as an agent focused on driving demand. She also explains the broader approach: structure teams to master the technology and abstract away choices like which model or tool to use, so customers can “plug in Toast and it works.”
- •Carlos characterizes the agent as handling marketing workflows across channels; Maggie names it Toast IQ Grow
- •Maggie frames the goal as driving demand—“put butts in seats”—and offloading that work
- •Carlos raises how restaurants show up in search and LLMs; Maggie responds that teams should handle that complexity for customers
- •Maggie says customers shouldn’t have to choose between providers or model types; Toast aims to handle it
- 20:31 – 21:54
Leadership in the AI era: staying hands-on with how teams build
Maggie explains why product leaders must remain close to execution as building practices change with AI. She shares how she stays in the weeds through direct management choices and temporary IC-style coverage, enabled by AI tools that let leaders scale differently.
- •Carlos asks how Maggie balances leadership with staying connected to customers and IC work
- •Maggie argues leaders must keep a hand in building or risk becoming irrelevant within months/years
- •She describes jumping into PM roles during leave or openings to stay close to frontline work
- •Maggie explains she manages some frontline teams directly and believes AI tools enable flatter team structures
- 21:54 – 30:45
Live demo: using Toast IQ to analyze performance and take action in plain language
Maggie demos Toast IQ in a realistic workflow: asking operational questions, finding drivers of performance, and making a menu price change through chat. The demo highlights faster access to insights, action execution (not just analysis), and a surprising early lesson that less tech-savvy operators adopted it quickly.
- •Maggie demonstrates a common operator flow: understand today/yesterday performance and compare periods
- •She shows querying drivers like average check and identifying items contributing to results
- •Maggie notes a key early learning: less tech-savvy operators loved natural language because they didn’t need to learn software
- •She changes a menu item price via chat and explains the interface generates action cards dynamically based on the request
- •Carlos asks about accessing Toast data from external assistants; Maggie says focus is currently on getting work done in Toast and notes ongoing customer conversations
