How to build your self-improving software factorySave your spot

Zesty connects a mature AI toolchain into governed end-to-end workflows

Zesty is a cloud optimization platform that automatically optimizes compute, storage, and Kubernetes resources based on real-time application demand. Zesty was already a mature, AI-enabled engineering organization, with AI capabilities across the software lifecycle. The gap was not in the individual tools or steps, but in connecting them into end-to-end workflows across systems.

Zesty logo
6 hours
Saved per engineer, per week
2 weeks
From initial setup to production
Industry
Cloud infrastructure optimization
Use case
Agentic SDLC orchestration: code review, PR descriptions, feature-flag cleanup, CI remediation
Integrations
Git source control, ticket and issue tracking, and communication channel systems, connected via native integrations and MCPs
Capabilities
Ticket-to-PR workflows, RBAC, full audit trail, human-in-the-loop approval, agent memory
The quality of the code review is fantastic, and developers miss it when it's not there. Developers are asking us to connect their repositories - that is how you know it's giving real value.
Dvir Cooper,R&D Group Manager, Zesty

The gap

Zesty's engineers were already using AI across the development lifecycle, with capable tools supporting individual stages. The missing layer was orchestration between those stages. A single unit of work could move across several systems, but no tool connected the full sequence end to end.

Zesty also required explicit permissions, a record of agent activity, and a human on every merge.

The transformation

With Overcut, Zesty connects these AI-enabled capabilities into governed workflows across systems, from ticket to merge pull request, with a full audit trail and developer approval at the final merge.

The first workflow, code review on GitLab merge requests, reached production two weeks after initial setup. Overcut was configured to Zesty's standards, including consolidating related findings, running checks in parallel, and tuning prompts against Zesty's own merge requests.

The principles applied with Overcut

  • Seamless AI Integration. Zesty was already using AI across the SDLC. Overcut connects those capabilities rather than replacing them.
  • Scale. Workflows run across repositories and teams, extending AI beyond individual developer interactions.
  • Governance. Explicit permissions, audit trails, and human approval are built into the platform.
  • Standards. Workflows are configurable to Zesty's engineering standards and can be continuously tuned as those standards evolve.

In practice: feature-flag cleanup workflow

The feature-flag workflow now runs on a schedule with nobody starting it. It identifies fully rolled-out flags, checks open merge requests and existing tickets to avoid duplicate work, opens a ticket, removes the retired code path, and raises a pull request. It then stops for a developer to review and merge.

Each individual step was already supported. Overcut connects the sequence across systems into one governed workflow.

After two weeks that the flag is GA'd, I want it merged automatically. This is why Overcut is perfect for it.
Dvir Cooper,R&D Group Manager, Zesty

Extending governance across the business

The governance model established with Overcut is now the role model for the wider organization. Zesty's R&D teams have adopted AI in their own domains, and bringing those workflows onto Overcut extends the same permissions, audit trail, and human-approval boundary that engineering already relies on - one control plane across every agent in the organization.

The race has
already begun

will you watch it, or will you win it?