Why We Invested in Pebble

TL;DR: Support tickets that require engineering attention are one of the biggest drains on developer time: they’re repetitive, unpredictable, and hard to delegate. Pebble is building an AI-powered platform that handles these tickets end-to-end: diagnosing, fixing, testing, and documenting bugs directly inside your existing workflow. We’re proud to back Alex, Dima, and the Pebble team as they help engineering teams reclaim thousands of hours spent on maintenance and support.

The problem: engineering time lost in support work

For fast-growing software companies, the backlog of support tickets that reach engineering teams can quickly spiral out of control.

Every week, engineers are pulled away from critical feature work to debug production issues, handle customer-reported bugs, and write tedious documentation - work that’s necessary but deeply unscalable.

A recent Stripe study estimates that developers spend over 25% of their time fixing bugs and maintaining code, and the Cost of Software Quality Report estimates that in the U.S. alone, companies spend over $600B annually finding and fixing bugs.

Existing AI tools (like Copilot or Cursor) help write new code, but not maintain existing systems. The real bottleneck isn’t code generation, it’s everything around it: triage, diagnosis, fixes, and follow-through.

What Pebble does

Pebble is an AI-powered software development platform purpose-built for engineering teams dealing with support and maintenance tickets.

It integrates directly into tools that engineering teams use like GitHub, Jira, Linear, Zendesk, and Sentry to:

• Diagnose customer-reported bugs from ticket data.
• Suggest and apply code fixes autonomously.
• Test the fix in a real environment.
• Generate release notes and documentation automatically.
  
  

Importantly, engineers stay in control. Pebble creates pull requests, runs tests, and explains every change before merging. One pilot customer saw Pebble complete a support task in under an hour that would’ve taken a senior engineer two to three days.

Pebble’s initial focus is fast-growing SaaS teams with high support volume (100+ tickets/month) and established AI budgets, where the time savings are immediate and measurable.

Why we’re excited

1. ‍ A sharp initial wedge with clear ROI
   Pebble’s first use case, automating support tickets that require engineering, is a sweet spot: painful enough to demand a solution, measurable in both time and dollars, and broad enough to serve as an entry point into the wider engineering workflow.
   ‍
2. A product that fits into how teams already work
   Instead of introducing another standalone AI tool, Pebble plugs into the systems engineers already use: from ticketing and CI/CD to observability. That makes adoption seamless for teams that can’t afford process disruption.
   ‍‍
3. Founders who deeply understand the pain
   Co-founders Alex Batchelor (PhD in Computational Neuroscience from Harvard, former Head of Engineering & AI at Mutiny) and Dima Ayyash (MIT-trained designer and UX lead) bring complementary perspectives, technical depth and design empathy. They’re building Pebble around the real workflows and pain points of software teams, not abstract AI research.
   ‍‍
4. Strong early traction and learning velocity
   Pebble is already running pilots and have 30+ companies in their pipeline, with rapid expansion driven by referrals. Early design partners have reported major time savings and high engagement, and the team expects to convert two-thirds of warm leads into paying customers.
   ‍‍
5. A roadmap to expand across the SDLC
   The long-term vision extends beyond support tickets: Pebble plans to orchestrate specialized AI agents across the entire software development lifecycle: planning, execution, QA, and release, becoming the connective tissue between teams, tools, and AI-powered software development.

Why now

Engineering teams are facing increasing complexity, shrinking budgets, and rising customer expectations. AI tools have made writing code easier, but maintaining production systems remains just as painful.

Pebble’s focus on the support-to-engineering bottleneck gives it a powerful early entry point, one that naturally expands into a broader platform opportunity as companies seek AI-native ways to manage their engineering operations.

Our thesis fit

At Propeller, we back the AI infrastructure and developer platforms reshaping how software gets built and maintained. Pebble embodies this vision: a pragmatic, high-leverage AI layer that doesn’t just help engineers code faster, but helps teams run better.

We’re thrilled to support Alex, Dima, and the Pebble team as they help engineering organizations turn reactive support into proactive, automated excellence.