MLOps.community
MLOps.community

Inside Just Eat's AI Lab: Voice Agents & Agentic Commerce

26 May 2026 1:18:38 Demetrios

Listen to episode

About this episode

Guthrie Cooper leads product for the AI Innovation & Incubation domain at Just Eat Takeaway, the company that launched Europe's first online food ordering platform back in 2000 and is now part of Prosus. He joins an engineering lead from the same team to break down how they run an "incubator within an incubator," shipping prototypes in 2 to 4 weeks and betting on where agentic commerce, voice, and physical AI are heading next.


We get into voice agents built on the OpenAI Realtime API, the "beyond the app" strategy of meeting customers inside ChatGPT, Claude, and Gemini, and what it actually takes to ship AI features inside a 25-year-old, 16-country company.


What we cover:


- Incubator within an incubator: How a pre-seed team turns brand-new protocols into working prototypes in 2 to 4 weeks.

- Beyond the app: Why Just Eat is building for ChatGPT, Gemini, and search instead of only defending its own app.

- Voice as a mode, not a feature: Interruptible, contextual voice agents, and why a "voice companion" is harder than a voice mode.

- Conversational continuity: Keeping context alive across car, phone, and laptop without breaking the flow.

- Agentic commerce and MCP: System-to-system and agent-to-agent integrations, plus the privacy and consent line they refuse to cross.

- Convenience vs grievance: Why proactive notifications are a churn risk, and how to nudge without getting muted.

- Context engineering for voice: Preloading context, parallel tool calls, caching menus, and keeping prompts light so the agent does not hallucinate.

- How they decide what to build: Lean startup stage gates, scanning the horizon, painted doors, and killing your darlings.

- Ideas that died: Why location-based collection notifications never shipped, and what conversational search accidentally fixed.

- Physical AI: Drone delivery with Manna in Ireland, RIVR robot dogs in Switzerland, and orchestrated last-mile handoffs between humans and robots.

- AI-first org: AI champions, PMs shipping code, evals, and scaling usability testing with synthetic personas.


If you build AI products, work in commerce or logistics, or care about how voice and agents reshape how people buy, this one is for you.


Links & Resources:

- Just Eat Takeaway: https://www.justeattakeaway.com

- Just Eat AI Voice Assistant announcement: https://newsroom.justeattakeaway.com/en-WW/259937-introducing-the-next-evolution-of-ordering-just-eat-takeaway-com-unveils-ai-voice-assistant/

- Manna drone delivery: https://www.manna.aero

- OpenAI Agentic Commerce Protocol: https://developers.openai.com/commerce

- Model Context Protocol (MCP): https://modelcontextprotocol.io


Timestamps

00:00 Meet the AI Innovation & Incubation team

00:30 The incubator within an incubator

02:30 Just Eat's history and the Prosus era

05:00 Beyond the app: ChatGPT vs Google search

07:15 Conversational continuity across devices

09:35 Sharing data with ChatGPT and Claude, and the privacy line

11:18 Health data, Whoop, and the proactive coffee order

13:10 Friction points are why the incubator exists

14:15 Agentic commerce, MCP, and agent-to-agent

20:30 Hyper-personalization and the personal shopper

21:50 Convenience vs grievance: the notification trap

33:50 Super apps, WeChat, and autonomous agents

37:00 How they decide what to build: lean stage gates

48:45 The idea that died: location-based notifications

54:35 Putting ideas on ice: the Mercedes-Benz in-car app

58:30 The hard parts of building a voice agent

1:00:25 Latency, the OpenAI Realtime API, and preloading context

1:08:00 Physical AI: drones in Ireland and robot dogs in Switzerland

1:15:35 AI-first: champions, evals, and synthetic user testing


#AgenticCommerce #VoiceAI #DroneDelivery

Want to find AI jobs?

Join thousands of AI professionals finding their next opportunity

We respect your inbox. Unsubscribe at any time.

© 2026 MLOps.community . All rights reserved.

Common Questions

Frequently asked questions

Quick answers about how DevFound's AI matching, resumes, and referrals work.

DevFound's AI Copilot ingests your profile, goals, and live job data to deliver curated matches in seconds. Every match includes a resume variant, suggested referrals, and interview prep so you can act immediately. The more feedback you provide, the sharper the Copilot becomes.

AI-led job searches shrink the hours spent sifting through boards and formatting resumes. DevFound pairs automation with your personal outreach, so you reserve energy for interviews and negotiation. Traditional networking still matters, but AI gives you a lift before you even send a message.

Modern AI roles expect comfort with production-grade code, data fluency, and practical ML tooling. The strongest candidates pair deep technical chops with storytelling—translating model impact to product, GTM, and exec partners. Continuous learning keeps you ahead as stacks evolve.

DevFound rewards active seekers. Keep your profile fresh, respond to match quality prompts, and enable alerts so you never miss a role. The AI prioritizes companies and teams that align with your feedback, accelerating both introductions and interview invites.

High-density tech hubs continue to host the deepest AI talent pools, yet distributed teams are catching up fast. Use DevFound filters to hone in on onsite, hybrid, or fully remote roles and watch openings expand across time zones.

DevFound aggregates thousands of remote AI openings and flags the nuances—core hours, async culture, and visa needs—up front. The Copilot also recommends how to position your distributed work experience so hiring managers know you can thrive on a remote team.