FAQs

What does petals.ai do?

Critical IT applications are hard to support. Companies often lack the documentation and the people who know how these systems work, and manual support is costly.

Petals.ai provides Iris, an AI-first IT application support agent that acts as a subject matter expert on your systems. Iris learns from:

  • Your application code
  • Pointed conversations with your experts
  • Your available documentation

When should you engage with petals.ai?

When you run a critical IT application that needs regular maintenance and business support, and you are:

  • At risk of losing the people who know the system
  • Working with low-cost offshore staff who may not be available right away
  • Paying too much to support the system manually

How does petals.ai work?

  1. Iris connects to your source code repository and documentation to understand the business logic.
  2. Your experts teach Iris what isn't written down.
  3. Users and support teams bring Iris issues at L1, L2 and L3.
  4. Iris diagnoses and recommends, and your people make the fix.

See how it works

How does Iris help at each support level?

  • L1, simple questions: users ask Iris directly, without raising a ticket.
  • L2, diagnose and fix: Iris finds the root cause using its database of root cause patterns, then gives your support team step-by-step instructions.
  • L3, features and bugs: engineers get the source code context and the applications affected before days of work begin.

Does Iris change our code or systems?

No. Iris diagnoses and recommends, and your people make the fix.

Do we have to replace our critical IT applications?

No. Iris is built to support the applications you run today.

Does this replace our offshore team?

It makes them faster, and keeps support moving when they aren't available.

Why can't I use ChatGPT or other LLMs myself?

  • The code base of a critical IT application can be millions of lines of code, which is too large to feed to an LLM.
  • You can't teach ChatGPT your company-specific business knowledge the way you can teach Iris.
  • Code has dependencies, which ChatGPT can't capture.
  • You would need a significant amount of prompt engineering to get the results you want.

About Petals.ai

Petals.ai is a product company based in Palo Alto, CA. We are a team of engineers and entrepreneurs with deep expertise in AI and critical IT applications.

We focus on MedTech, Manufacturing and Retail.