Select Page

Gopal Dommety

|
originally published on Dec 19, 2025
Share

When we started working deeply with security teams and developers, one thing became obvious very quickly:

Finding security problems is no longer the bottleneck.
Fixing them is.

Most organizations today already run a solid set of scanners:

  • SAST for code
  • SCA for dependencies
  • IaC and container scanners
  • Cloud and runtime security tools

Detection isn’t the issue.
The industry has largely solved “what’s wrong.”

What remains painfully unsolved is:

  • What actually matters?
  • How do we fix it safely?
  • How do we do it without slowing developers down?

AI Guardian exists because we keep seeing the same failure pattern over and over again.

The Reality We Observed (Across Dozens of Teams)

Security issues don’t fail because people don’t care.
They fail because remediation doesn’t fit how software is built.

Here’s what actually happens in most organizations:

  • A scanner detects an issue
  • A ticket or PR is created
  • A developer is asked to “fix security”
  • The developer stares at:
    • A CVE
    • A rule ID
    • A vague recommendation
  • The real work begins:
    • Understanding context
    • Figuring out blast radius
    • Making sure functionality doesn’t break
    • Asking a senior engineer or AppSec for help

This is where things slow down.

The harder the issue, the more it escalates:

  • From junior dev → senior dev
  • From dev → AppSec
  • From AppSec → architecture discussions
  • From architecture → delayed releases

The issue isn’t awareness.
The issue is remediation complexity.

The Three Buckets That Consume Developer Time

From a technical standpoint, nearly all remediation work falls into three buckets:

1. Code vulnerabilities

  • Secrets
  • Injection flaws
  • Unsafe APIs
  • Access control mistakes
  • Crypto and error-handling issues

These are frequent and noisy.

2. Dependency vulnerabilities

  • Direct dependencies
  • Transitive dependencies
  • Forced upgrades due to CVEs
  • Breaking API changes

These are fewer, but much more disruptive.

3. Infrastructure and configuration issues

  • Dockerfiles running as root
  • Insecure base images
  • IaC misconfigurations
  • Cloud permission problems

These are often copied, repeated, and poorly understood.

Most tools can detect all three.
Almost none can remediate all three safely.

Where Existing Tools Break Down

We looked closely at the tools teams already use.

  • Dependency bots can bump versions, but they don’t refactor code.
  • Autofix features can rewrite patterns, but they don’t understand application behavior.
  • Cloud tools can flag misconfigurations, but they don’t operate in developer workflows.

And critically:

None of these tools understand the system as a whole.

They don’t know:

  • How code, dependencies, artifacts, and infrastructure connect
  • Which execution paths matter
  • What context is required to fix this issue, in this system, right now

That missing context is where humans step in—and where time disappears.

The Insight That Led to AI Guardian

The breakthrough insight for us was simple:

Remediation is not a single problem.
It’s a set of domain-specific problems that require shared context.

Fixing a SQL injection is not the same as:

  • Fixing a transitive dependency
  • Fixing a Docker misconfiguration
  • Fixing a runtime-only vulnerability

But they all require:

  • System context
  • Dependency awareness
  • CI/CD and deployment understanding
  • Confidence that functionality won’t break

So instead of building another scanner, we built AI Guardian around remediation.

What AI Guardian Actually Does

AI Guardian is designed to sit after detection and inside the developer workflow.

Conceptually, it works like this:
1.  Ingest findings
From any scanner—ours or third-party.
2. Correlate context
Using data from:

  • Git
  • CI/CD
  • Artifacts
  • Infrastructure
  • Cloud and runtime signals

3. Prioritize what matters
Not every CVE deserves the same response.
4. Invoke specialized remediation agents

  • Code agents
  • Dependency and artifact agents
  • IaC and Docker agents
  • Cloud and runtime agents

5. Remediate where developers work

  • IDE
  • GitHub PRs
  • CI/CD pipelines

6. Verify both security and functionality
A fix that breaks production is not a fix.

This is not about “AI suggestions.”
It’s about producing correct, reviewable changes.

Why Context Is the Hard Part

Many of the hardest vulnerabilities cannot be fixed by looking at:

  • A CVE description
  • A single file
  • A static rule match

Correct remediation often requires understanding:

  • How the application is deployed
  • Which dependencies are actually reachable
  • Which environment (dev vs prod) is at risk
  • Where authentication and authorization are enforced
  • How infrastructure and runtime behavior interact

This is why so many “easy fixes” turn into multi-day efforts.

AI Guardian is built on the premise that context is the input, not an afterthought.

What Changes When Remediation Becomes Developer-Native

When remediation:

  • Happens automatically
  • Is context-aware
  • Lives inside Git and CI/CD

We see a real shift:

  • Junior developers can fix issues they previously avoided
  • Senior engineers stop getting pulled into fire drills
  • AppSec teams move from ticket triage to risk ownership
  • Vulnerabilities close faster, with fewer regressions

Security stops being a tax.
It becomes part of normal development flow.

Why We Care About This Problem

We didn’t build AI Guardian to replace developers or AppSec teams.

We built it because:

  • We’ve watched teams lose weeks to a single dependency upgrade
  • We’ve seen security backlogs grow despite “best-in-class” tooling
  • We’ve seen good engineers burned out by repetitive, high-risk fixes

The industry doesn’t need more alerts.
It needs better outcomes.

AI Guardian exists to make fixing security issues:

  • Faster
  • Safer
  • And aligned with how software is actually built

That’s the problem we set out to solve with AI Guardian.

Gopal Dommety, Ph.D. is the Chief Executive Officer of OpsMx, a company advancing the automation and security of software delivery for the modern enterprise. Under his leadership, OpsMx is redefining how organizations build, secure, and release software, enabling developers to deliver innovation with speed, safety, and confidence. A technologist and inventor, Dr. Dommety holds over 70 patents and is the principal author of several Internet Protocols (RFCs) that power today’s global networking infrastructure. His work has shaped critical areas of large-scale distributed systems, algorithmic design, and secure automation. He has also authored more than 20 peer-reviewed papers, book chapters, and journal publications, and previously led the Mind-Map Project, an AI research initiative focused on modeling behavioral and personality traits from user-generated data. Before founding OpsMx, he was a General Partner at Neem Capital, a technology-focused investment firm, and held senior leadership roles in product management, research, and engineering at major technology companies and startups. Rooted in humble beginnings from a remote village in India, Gopal’’s career is guided by the principles of simplicity, first-principles thinking, and purpose-driven innovation—values that continue to shape his vision for building secure, intelligent, and resilient technology systems that move the world forward.

0 Comments

Submit a Comment

Your email address will not be published.

This site uses Akismet to reduce spam. Learn how your comment data is processed.