Building Lorikeet: How AI Humility and a Dual-Agent Architecture Are Redefining Customer Support

Building Lorikeet: How AI Humility and a Dual-Agent Architecture Are Redefining Customer Support

Guests: * Jamie Hall, Co-founder & CTO, Lorikeet * Xharmagne Carandang, Product Engineer, Lorikeet * Rona Wang, Product Engineer, Lorikeet In this episode: * How Lorikeet evolved from failed ops tools to a full AI customer support concierge * The dual-agent architecture: Concierge for customer tickets, Coach for configuration and ongoing improvement * Why "AI humility" — defaulting to human handoff when uncertain — is a core design principle * How Lorikeet integrates with Zendesk and Intercom instead of replacing them * The UX evolution from workflow builder to conversational interface — and why the blank chat box is still hard * "Resolution in the loop": how human agents unblock the AI without taking over a ticket * Why guardrails need to be domain-specific — the cannabis company story * How customers define their own evals and guardrails through the Coach interface * Using AI to diagnose failure modes in traces and automatically suggest fixes * Lorikeet's product engineering culture: every engineer asks weekly what they learned from a customer Resources & Links: * Lorikeet [https://www.lorikeetcx.ai/?ref=producttalk.org] — AI customer support concierge for enterprises in regulated industries * Gradient Labs on Just Now Possible [https://www.producttalk.org/building-a-multi-agent-platform-with-gradient-labs/?ref=producttalk.org] — another AI agent team in regulated financial services * Neople on Just Now Possible [https://www.producttalk.org/building-ai-coworkers-how-neople-is-making-agents-work-where-you-work/?ref=producttalk.org] — AI digital coworkers with a similar training-by-conversation approach * Incident.io on Just Now Possible [https://www.producttalk.org/when-ai-becomes-your-sre-how-incident-io-is-automating-incident-response/?ref=producttalk.org] — AI SRE with multi-agent hypothesis investigation Chapters 00:00 Meet the Team 01:05 What Lorikeet Builds 02:34 Origin Story and Early Missteps 06:42 Finding the Real Support Pain 07:37 Why AI Fits Support Work 11:16 First Prototype and Early Evals 14:42 Design Partners and Selling the CLI 16:30 Product Mindset and the Real Moat 19:47 Rona Joins and Scaling Up 21:02 Milestones Voice Actions Escalation 23:48 Integrations with Zendesk Intercom 25:59 How the Agent Works Today 28:30 Coach Agent and Configuration UX 32:58 SOPs to Test Cases 34:35 Refund Flow Setup 36:12 Coach Conversational UI 38:12 Hybrid UX Guidance 40:46 Resolution in Loop 43:17 Collaboration Middle Ground 49:40 Process Maturity Limits 53:30 Confidence and Guardrails 55:59 Customer Defined Guardrails 01:01:14 Trace Diagnosis Agent 01:03:14 Product Engineers Culture 01:07:46 Closing Thoughts

Building Rhea's Factory: How AI-Designed Enzymes Could Finally Solve Plastic Recycling

Guests * Arzu Sandıkçı, Co-founder & CEO, Rhea's Factory * Mert Topcu, Co-founder, Rhea's Factory In this episode: * Why only 10% of plastic gets recycled—and why mechanical and chemical methods hit a ceiling * How enzymatic recycling breaks plastic all the way back to its original monomers, unlike traditional methods that just shorten polymer chains * Why enzymes are selective: they can target specific plastic types even in mixed waste streams * The discovery of a plastic-eating bacteria in Japan that opened the door to enzymatic recycling * How AlphaFold and the Nobel Prize in Chemistry transformed what's possible in enzyme engineering * How Rhea's Factory uses protein language models (PLMs) and multi-step AI pipelines to design novel enzymes computationally * The evolution from a human-orchestrated pipeline to an agentic AI scientist * How guardrails at each pipeline step keep the AI pointed in the right direction without limiting exploration * Why wet lab data—even just hundreds of proprietary data points—can be enough to train a powerful domain-specific prediction model * Why Mert sometimes wants the model to hallucinate (and how high temperature settings help explore the full enzyme design space) * The business constraint: enzymatic recycling must compete economically with cheap, oil-based plastic production * What's next: a process agent, a 5,000-ton demo plant in California, and enzymes for new plastic types Resources & Links * Rhea's Factory [https://rheasfactory.com/?ref=producttalk.org] — Enzymatic plastic recycling technology * AlphaFold [https://alphafold.ebi.ac.uk/?ref=producttalk.org] — DeepMind's AI system for protein structure prediction (inspiration for the Nobel Prize in Chemistry) * Maven AI Evals Course [https://maven.com/parlance-labs/evals?promoCode=torres-35&ref=producttalk.org] — The course Teresa took to learn about evals (35% off with Teresa's affiliate link) Chapters 00:00 Meet the Founders 01:50 Why Plastic Circularity 03:19 Mechanical vs True Recycling 04:52 Biology as the New Tool 07:20 Necklace and Pearls Analogy 13:22 Low Energy Reactor Process 17:33 Origin Story and PET Enzyme 22:52 Protein Folding and AlphaFold 28:32 AI Designed Enzymes 34:28 Protein Language Models Stack 37:14 Multi Step Protein Generation 39:00 Building on Foundation Models 40:50 Lab First Success Metrics 43:10 From Human to Agentic Orchestration 43:59 Problem Statements as Inputs 46:18 Guardrails at Every Stage 47:48 Prediction Models and Data Limits 50:03 Industrial Reality and Cost 52:30 Agentic Parallels and Orchestrators 57:45 Impact on Timelines and Diversity 01:03:23 When Hallucination Helps 01:04:09 Scaling Up and Process Agents 01:06:56 Enzyme Blends for Mixed Plastics 01:07:49 Why Clamshells Aren't Recyclable 01:09:34 Closing Thoughts and Thanks

Building AI Employees for Hospitality: How AITropos Takes Orders Where Customers Already Are

Guests * Santi Marchiori, CEO, AITropos * Juan Haedo, CTO, AITropos You'll hear how they * Spent two years exploring hundreds of startup ideas before finding the specific niche of AI-powered order taking in hospitality * Went through three product iterations — hardware for waiters, a waiter app, and finally a customer-facing WhatsApp agent — before landing on the right form factor * Identified order item identification accuracy as their single most important KPI * Chose a tools-based agent architecture over MCP or pipelines to hit real-time response speed requirements * Built a parallelized pipeline that searches for multiple products simultaneously and pre-fetches product context before the agent even calls a tool * Use smaller, fast sub-agents to build an "immediate system prompt" that injects relevant data into each turn without extra tool calls * Test with thousands of agent-simulated customer conversations run overnight before deploying to new venues * Reduced new customer onboarding from three months to a few weeks — and continue to shrink it as they build domain templates Resources & Links * AITropos [https://aitropos.com?ref=producttalk.org] Chapters: 00:00 Meet the Founders 00:59 What Tropos Builds 01:51 AI vs Human Touch 06:17 Restaurant Use Cases 08:16 Why Hospitality 10:47 Finding the Wedge 16:00 Early Prototypes 16:46 Hard Parts of Ordering 18:03 Speed and Channels 21:15 Iteration and Model Jumps 30:50 Customer Order Flow 35:48 Menu Discovery Question 36:07 Menus Inside WhatsApp 36:50 Finding the Chat Entry 37:37 Why Text Ordering Wins 38:30 Under the Hood Pipeline 40:54 Tools Over Workflows 45:05 Tooling and Prompt Composer 49:29 Preloading Context Fast 54:02 Founder Learning Mindset 57:21 Evaluating Order Accuracy 01:00:03 Testing and Human Takeover 01:03:56 Onboarding and Scaling Up 01:06:10 Whats Next and Wrap

Building Todoist Ramble: How Doist Turned Voice Braindumps into Real-Time Task Capture

Guests * Ernesto Garcia, Front-end Product Engineer, Doist * Thomas Jost, Backend Software Engineer, Doist * Hugo Fauquenoi, Product Manager, Doist In this episode * How Doist's 2-3 month AI exploration phase led to Ramble — and why voice-to-task emerged as the top contender * The user research insight behind Ramble: people using pen and paper or ChatGPT voice to brainstorm tasks before committing them to Todoist * Why Ramble skips transcription entirely and processes raw audio directly with a Gemini live audio model * How the model makes tool calls (add task, edit task, delete task) in real time while the user is still speaking — no text output at all * Designing for the driving use case: sound effects as audio confirmation cues alongside visual task cards * The challenge of teaching an LLM to capture tasks literally without over-interpreting or doing them — and how temperature tuning played a role * Date handling complexity: injecting the current date, normalizing to days vs. months, and always outputting dates in English for the natural language parser * Building an LLM-judge eval system with 20+ language recordings from 100+ employees across 35 countries to catch prompt regressions * Why Doist chose to inject the full project/label list into the system prompt instead of building a RAG pipeline — and why it worked * How easy correction beats perfect first-time accuracy in natural language interfaces * What's next: multimodal task capture from images and text blobs, Apple Watch support, and automation integrations Resources & Links * Todoist [https://todoist.com/?ref=producttalk.org] * Doist [https://doist.com/?ref=producttalk.org] * Google Vertex AI (Gemini) [https://cloud.google.com/vertex-ai?ref=producttalk.org] Chapters: 00:00 Meet the Doist Team 01:40 What Doist Builds 02:27 Ramble Voice to Tasks 04:16 Why Voice Matters 07:42 Brain Dump Insight 09:46 Prototyping With LLMs 11:08 Live Audio Workflow 14:32 Driving Friendly UX 18:47 Tool Only Architecture 26:06 Evals and Multilingual Testing 28:41 Taming Dates and Time 33:28 Fixing Date Confusion 33:43 Defining Task Boundaries 34:34 Capture Versus Do 37:17 Tuning Creativity Levels 39:01 Evals Across Languages 41:23 Feedback and Regressions 44:09 Model Upgrades Over Time 46:33 Projects Labels Context 51:40 Handling Ambiguous Names 54:23 Whats Next Multimodal 58:48 From Capture to Execution 59:46 Closing Thoughts

Building Banani: How a Canvas-First AI Designer Is Raising the Floor on Product Design

Guests * Vlad Solomakha, CEO & Co-founder, Banani * Vova Parkhomchuk, CTO & Co-founder, Banani * Vlad Ostapovats, Founding Growth, Banani In this episode * Why Banani started as a Figma plugin and what they learned from early organic distribution * The canvas-first approach: why Banani is built around a design canvas rather than a chat interface * How their agent architecture splits prompts into surgical edits instead of regenerating full screens * The "gulf of specification" problem and what Banani is building to help agents and designers speak the same visual language * Managing context across canvases with hundreds of screens — per-screen history with shared project context * Why Banani doesn't compile running applications — just HTML/CSS mockups — and how that shapes everything * How they evaluate design quality without traditional evals: spinning up 10 screens from one prompt to compare models * Their approach to building at the edge of what's possible: identifying which model limitations to work around vs. wait out * The role of context engineering and specialized agent tools in producing tasteful, high-quality design Resources & Links * Banani [https://www.banani.co/?ref=producttalk.org] * TL Draw [https://tldraw.com?ref=producttalk.org] CHAPTERS 00:00 Meet the Founders 01:12 What Bonani Builds 02:18 Why an AI Designer 03:40 Raising the Design Floor 06:23 Why AI Was Finally Ready 10:48 First Prototype Figma Plugin 14:10 Early Growth and Distribution 15:25 Standing Out in a Crowded Market 20:13 Product Tour Canvas First AI 23:40 Autopilot vs Manual Control 27:07 Tech Behind High Quality Design 32:08 Craft Beyond 80 Percent 33:40 Gulf of Specification 36:44 Proactive Agent Interviews 38:40 Canvas First UX Choices 42:54 Agent Architecture Under Hood 48:48 State History Context Tricks 52:32 Tooling Context Engineering 56:04 Navigating Busy Canvases 01:00:13 Betting on Model Progress 01:03:47 Shipping Around Imperfections 01:07:20 Try Banani and Next Steps 01:07:52 Building the Banani MCP 01:09:19 Final Thanks and Wrap