CHAPTERS
Yhangry’s AI agenda: three practical business use cases
Siddhi Mittal frames the talk around three concrete ways Yhangry is applying AI agents across the company. She positions these as high-leverage, ROI-driven initiatives tied to growth and operational scale.
Engineering pain point: bug backlog vs. higher-ROI roadmap work
She explains how small but annoying bugs get deprioritized when the team is chasing bigger growth initiatives. With $15M GMV and a goal to 10x, engineering time is too valuable to spend on low-priority fixes.
Building an autonomous bug fixer in under four days
During maternity leave, Siddhi builds an autonomous bug-fixing agent quickly and puts it into production. The result is immediate: dozens of bugs fixed and shipped in the first week.
Quality + context: benchmark pass rates and self-improvement challenge
She shares rough performance expectations for one-shot bug fixing and highlights the main limitation: insufficient context. The key learning is that agent performance hinges on feeding the right system and codebase context so it can improve over time.
Growth hack: teaching AI agents in plain English as a distribution channel
Siddhi discovers that explaining AI agents simply is a rare, valuable skill and turns it into a marketing wedge. Instead of pitching Yhangry directly, she offers conference talks on building agents, which conference organizers eagerly accept.
Turning conference talks into product demos and partnership leverage
She uses the speaking slots to seamlessly incorporate Yhangry’s affiliate/partner angle and demo the new AI product. Attendees share screenshots and post about it afterward, amplifying reach on LinkedIn.
Why the current booking flow feels ‘old school’ and leaky
Siddhi explains the core product friction: booking a chef requires too many steps and too much back-and-forth. This causes conversion loss despite Yhangry having enough historical data to streamline matching.
Yhangry AI on the customer side: instant, higher-confidence chef matching
She outlines the customer-facing vision: use Yhangry’s data to deliver fast, accurate, one-shot matches. The goal is to eliminate repetitive communication and accelerate time-to-booking.
Yhangry AI on the chef side: ‘Claude for chefs’ to remove admin burden
On the supply side, chefs waste time repeating the same information to many customers. The proposed agent acts as a chef’s assistant, handling repetitive admin and communications to free chefs for cooking.
MVP reality check: chef needs are highly variable
She notes the hardest product challenge: chefs are not a uniform user group. Their workflows and preferences diverge significantly, making it difficult to define a minimal version that’s reliable enough to ship.
Company-wide shift: making every workflow AI-native
Siddhi states that agents are spreading across all functions in the company. The strategy is to rebuild processes with AI in mind rather than bolting tools onto old workflows.
Founder acceleration and hard org calls: tools obsession and leadership changes
She recounts how personally adopting AI tools intensified her pace and standards, leading to major org decisions. She describes firing a tech lead who became a skills ceiling and hiring a new head of engineering quickly.
Operationalizing learning: weekly agentic labs and shared understanding
Because people’s agent-building abilities vary, the team creates weekly labs to align skill levels and quality. They use Claude/code tooling to translate discussions and transcripts into diagrams so everyone can follow.
Founder brand as a growth engine: give away expertise, funnel attention back
Siddhi closes by articulating her take on founder brand: do what you’re uniquely good at publicly, earn attention, and route it back to the company. Her unique edge is translating deep AI knowledge into simple language and distributing it freely.
Get more out of YouTube videos.
High quality summaries for YouTube videos. Accurate transcripts to search & find moments. Powered by ChatGPT & Claude AI.
Add to Chrome