At Rabbit, we’ve always thrived on speed, from our dark stores to delivery times, velocity is our edge. But in early 2025, we found ourselves facing a profound realization: the way we were building software didn’t match the speed at which our business was evolving. JIRA tickets, grooming sessions, sprint planning, manual QA cycles , all of it was dragging us down while the world around us moved at breakneck pace.

So we asked ourselves a simple but transformative question: What if we reimagined everything?

The Spark: Rethinking Work with AI

AI wasn’t new to us. Rabbit had already integrated AI into demand forecasting, route optimization, and customer care ..etc. But what about the way we work internally , the way engineering, product, and business functions collaborate to bring ideas to life?

We believed we could unlock 10x delivery speed by trusting AI as a core team member, not just a tool in our toolkit.

🔁 From the Old Way… to the New Way of Work

The Traditional Model:

• Work organized in sprints with rigid ceremonies
• Requirements detailed in JIRA tickets
• Developers writing all code manually
• Manual QA processes and batch deployments
• Progress measured with story points
• Features taking 2-3 weeks to ship

Our New AI-Native Model:

• Continuous prompt-driven delivery replacing sprint cycles
• Structured prompt specs instead of JIRA tickets
• AI-assisted code generation using tools like Cursor and GPT
• Prompt-generated tests replacing manual QA
• Continuous deployment with feature flags and instant rollbacks
• Prompt throughput as our new velocity metric
• Features shipping in 1-2 days

We’re moving from building features to co-creating with AI. And the impact is immediate.

🧪 How We Made the Shift: The Pilot Program

We launched a one-month pilot with two squads (Sales & Operations) to test our hypothesis. Here’s exactly how we did it:

Structural Changes:

• Restructured into small cross-functional squads (3-4 people) owning the full cycle
• Adopted a Kanban model with daily ticket prioritization
• Each squad picks up new priority tickets daily

Process Elimination:

• Eliminated traditional ceremonies: grooming, sprint planning, regression testing
• Removed the overhead that was slowing us down

New Capabilities:

• Delivered training on prompt writing and AI tool usage
• Benchmarked velocity and quality week over week
• Built new rhythms for fast learning cycles without losing visibility

The result? More throughput, less overhead, and faster iteration.

🧠 Solving the New Challenges

As we shifted, we had to answer critical questions that didn’t exist in our old model:

How do we commit to release dates with prompt-driven flow? We now define tickets as small, medium, or large , decided by a mix of product and engineering input, giving us predictable delivery windows.

Who explains tickets? We created a two-week buffer where product and business stakeholders prepare and clarify new prompts, ensuring clarity without slowing momentum.

How do we maintain quality? By building quality into the AI-assisted development process rather than catching issues after the fact.

🛠️ Impact Beyond Engineering: Company-Wide Transformation

This transformation didn’t stop with the engineering team. It forced a mindset shift across all of Rabbit:

Product Teams now think in iterations, not perfection , they test quickly, learn fast, and refine. They’re writing prompt-based specs and working in tight feedback loops with engineering to deploy ideas in days, not weeks.

Business Teams have moved from long PRD documents to shaping ideas collaboratively, shifting from “big launches” to rapid experimentation , learning fast, iterating faster.

Marketing is using AI for personalization at scale, automated content generation, and performance optimization.

Customer Care is powered by AI copilots and bots, enhancing service speed and satisfaction.

Data and Operations are embracing AI for smarter decision-making, anomaly detection, and automated insights.

Cross-functional alignment is tighter than ever, with everyone thinking prompt-first and experimenting over perfecting.

🧬 The New Engineering Mindset

In this AI-native world, our engineers have transformed from builders of features to co-creators of business innovation.

They’re now:

• Co-pilots of ideas , partnering early with business teams
• Designers of prompts , crafting specifications that AI can execute
• Rapid testers of hypotheses , prototyping quickly with AI assistance
• Partners in shaping the business , not just implementing predetermined solutions

Engineers use AI to reduce cognitive load and repetitive tasks, allowing them to focus more on impact and less on process. They’re not waiting for specs , they’re co-owning the problem space, not just the solution.

We don’t just build what we’re told. We challenge ideas, iterate fast, and drive results.

🧭 What’s Next: Scaling the AI-Native Company

We’re scaled this model across all squads because we’ve learned something fundamental: AI isn’t just a tool , it’s now part of our DNA.

As we evolve, we’re continuously learning how to:

• Shape better prompts for more precise outcomes
• Automate more edge cases and complex scenarios
• Expand the model to lifecycle operations, support, and growth initiatives
• Build a company where every role is AI-native and every process is continuously evolving

To keep winning, every team at Rabbit is expected to ask: “How can AI help me do this better, faster, and with more impact?”

This isn’t about replacing people. It’s about amplifying our creativity, impact, and speed , with humans in the loop, empowered by AI.

The New Beginning

This isn’t the end of our transformation story. It’s our new beginning , and it’s faster, smarter, and more fun.

We’re building a company where velocity isn’t just about delivery speed, but about the speed of learning, adapting, and innovating. Where AI doesn’t replace human creativity but amplifies it exponentially.

Welcome to Rabbit’s New Way of Work.

We’re just getting started. 🐰