About the Role
Toast is launching a new initiative to bring the power of voice AI into the in-store restaurant experience. While voice has transformed drive-thru and phone ordering, there is a large opportunity to make in-store experiences better for guests and more proficient for restaurant teams. This team will explore voice capabilities across front-of-house (guest ordering, service interactions, POS assistance) and back-of-house (kitchen coordination, hands-free workflows, inventory signals). As the Principal Product Manager — Voice AI (In-Store) on the Commerce Platform, you’ll define strategy, evaluate vendors and partners, run zero→one discovery and pilots in real restaurants, and lead commercialization — creating a reusable voice platform that powers multiple use cases across the restaurant.
Responsibilities
- Define vision & portfolio strategy. Own the long-term vision and multi-quarter roadmap for in-store voice AI and the platform that enables multiple use cases across FOH/BOH. Balance short-term pilots with long-term platform investments.
- Run vendor & technical evaluation. Lead technical evaluation of voice and AI vendors (ASR, NLU, TTS, edge inference, on-device options) and create benchmarking frameworks for accuracy, latency, privacy, and cost.
- Lead commercial conversations. Drive vendor commercial discussions (pricing models, licensing, SLAs) and surface negotiation points — partnering with Procurement and Legal who finalize contracts and budget approvals.
- Drive zero → one discovery and pilots. Plan and execute discovery, rapid prototyping, and in-restaurant pilots; iterate based on operational feedback and scale successful experiments.
- Design a reusable voice platform. Define APIs, integration patterns, data flows, and tooling that make voice capabilities reusable across guest experiences, staff workflows, and integrations with POS, kitchen systems, and telemetry services.
- Lead GTM & commercialization. Define pricing, packaging, and go-to-market strategy with Product Marketing, Sales, and Operations; identify commercial models that fit restaurants of different sizes.
- Coordinate cross-functional delivery. Align Engineering, Hardware Ops, Reliability, Security, Data, Research, Design, Care/Onboarding, and GTM partners to de-risk delivery and adoption.
- Drive operational readiness & reliability. Define operational playbooks, observability, edge/online failover modes, and post-launch monitoring to ensure reliable in-store behavior and low support cost.
- Champion field research & validation. Spend time in restaurants to learn operational workflows, validate prototypes, and iterate with real staff and guests.
- Mentor & shape product practice. Coach senior PMs and engineers on voice/AI product development, vendor evaluation, and responsible AI practices; contribute to the Product Playbook.
- Embrace and advocate for the use of AI tools to accelerate product discovery, streamline execution, and explore new product experiences - helping shape how we build and what we build.
Requirements
- AI product delivery experience. ~8+ years shipping complex software products with demonstrated ownership of multi-team programs or a major product portfolio, and experience delivering AI-enabled features or products.
- AI Model Evaluation. Hands-on experience defining and owning AI model evaluation (examples: WER/CER, intent accuracy, wake-word false accept/reject, latency, completion/task success), building continuous evaluation and model ops processes (drift detection, automated regression tests, A/B testing, rollback/mitigation playbooks), and partnering with ML teams to run experiments and error analysis.
- Zero → one experience. Proven track record taking a product from discovery through prototype and pilot to scale.
- Vendor evaluation & commercial acumen. Hands-on experience evaluating AI/voice vendors and participating in vendor commercial conversations (note: Procurement/Legal finalize contracts).
- Platform mindset. Ability to design reusable APIs and platform capabilities that enable multiple downstream use cases and teams.
- Customer & ops empathy. Experience conducting on-site discovery and validating prototypes in operational environments (restaurant or other physical venues).
- Cross-team leadership. Exemplary stakeholder management and ability to align senior partners across hardware, engineering, ops, and GTM, especially on high-visibility initiatives.
- Communication & influence. Clear, persuasive written and verbal communication; able to synthesize technical tradeoffs and business value for executives and partners.
Nice to Haves
- Speech / voice product experience. Prior experience building voice AI products (ASR/NLU/TTS, or working directly with speech teams) is helpful but not required.
- Conversational UX & dialog design. Experience designing conversational flows, state management, turn-taking, error recovery and graceful degradation.
- Exposure to hardware considerations for voice (far-field audio, beamforming, noise cancellation, embedded audio processing).
- Experience with on-device / edge inference and hybrid cloud/edge voice architectures (model quantization, latency/footprint tradeoffs).
- Familiarity with privacy, compliance, and PII handling for voice data (on-device vs cloud processing, country-level regulations).
- Background in restaurant, retail, or other physical-world operations.
- Experience with ML lifecycle tooling, continuous model evaluation, and production A/B testing strategies for AI features.
Success Metrics
- Launch and adoption of successful in-store voice pilots (pilot → scale conversion).
- Measurable operational impact (e.g., staff time saved, reduction in order errors, throughput gains).
- Guest experience improvements (NPS/CSAT lift for voice flows, voice take rate).
- Platform reuse (number of distinct use cases / teams reusing core voice APIs).
- Reliability & support outcomes (uptime targets, reduction in Care tickets related to voice).
- AI health & model performance: model accuracy (WER/intent accuracy), latency, drift rate, and rate of regressions caught by CI/monitoring systems.