CodeWall Raises $4M Pre-Seed for Autonomous Security

CodeWall is hiring a technical customer success manager to stay with security teams from scoping through remediation and renewals. That job description puts a human assignment inside the London company's autonomous-security pitch: help a customer turn an attack finding into work that gets finished.

CodeWall announced a $4M pre-seed round on October 14, 2026, led by Osney Capital with participation from InReach Ventures, Antler and angel investors. Founder and CEO Paul Price is building autonomous attack agents for enterprise security, pursuing sustained external testing as organizations change their applications and defenses.

The funding supports broader enterprise coverage, testing and remediation workflows, and hiring across technical and commercial teams in London. For buyers, the interesting shift is the relationship with the assessment: evidence can be revisited after a repair, giving the security team another opportunity to check the path it thought it had closed.

What CodeWall's $4M pre-seed will support

The financing accompanies CodeWall's emergence from stealth and a development agenda that stretches beyond finding the next vulnerability. The company says it will extend where its agents can operate across the enterprise while developing sustained testing, verification and remediation workflows. Its announced hiring spans offensive security, engineering, customer success and go-to-market work.

Osney Capital invests exclusively in the early-stage UK cybersecurity ecosystem. That specialization gives the lead investor a clear relationship to CodeWall's market, where the customer must be comfortable letting software investigate weaknesses in a live business environment. The firm's public positioning emphasizes sector knowledge and help for founders during early growth.

The other named participants bring their own early-stage focus. InReach Ventures describes an investment approach centered on European AI applications with a working product and initial user engagement, while Antler invests from the earliest stages of company building. Their participation identifies the financing syndicate; the announcement leaves individual check sizes and valuation undisclosed. The useful information for an enterprise buyer remains what CodeWall plans to build and support with the capital.

How autonomous external testing changes the evidence

CodeWall describes an outside-in starting position, with no supplied credentials, source code or insider access. Its specialist agents investigate in parallel and share discoveries through a persistent investigation graph, allowing a weakness in one application to inform investigation elsewhere. The company's premise is that connected systems need connected evidence: an opening matters partly because of where it can lead.

That approach changes the question a customer can ask of a finding. A vulnerability ticket identifies something to fix; an evidenced path helps explain what could become reachable through a combination of weaknesses. For the person deciding which engineering work should interrupt an existing plan, the connection between exposure and consequence can make the discussion more concrete.

CodeWall's public research portfolio includes work concerning McKinsey, Boston Consulting Group and Bain. Those organizations are research subjects in the company's account, and should not be read as a customer list. The research provides a way to examine the company's methods and disclosure work without treating every company-reported result as a universal measure of product effectiveness.

The broader product proposition reaches past the first investigation. Applications change, new systems appear and repairs alter the environment, creating reasons to revisit earlier conclusions. A security team buying sustained testing is therefore also buying a continuing conversation about which evidence remains current. CodeWall has made that return visit central to its commercial identity.

The people behind the remediation workflow

The roles advertised by CodeWall make its delivery ambitions unusually legible. Alongside software and security-research positions, the company lists a technical customer success manager whose remit runs from onboarding and scoping through findings, remediation and renewals. That sequence connects the technical result to the customer's continued willingness to use the service.

A penetration tester or red teamer is asked to validate agent findings and feed human tradecraft into autonomous attacks. A forward-deployed offensive security engineer is meant to bring CodeWall into customer environments and help teams put autonomous offense to work. These are advertised responsibilities, rather than evidence that every role has already been filled, but they show the range of work the company expects around the product.

For a buyer, that surrounding work has practical value. Someone must agree what can be tested, understand what a finding means, get the right team involved and return to the affected system after a change. Faster discovery can put more pressure on those handoffs, particularly when the engineer responsible for a repair has other priorities and needs reproducible evidence before changing a production application.

Paul Price's financing gives the company resources to develop both the agents and the organization around them. The staffing plan puts customer understanding alongside offensive capability, tying the quality of an investigation to whether a security team can carry it into action. A repair that survives the next authorized investigation gives that relationship another piece of evidence to work with.

What enterprise buyers need to establish

CodeWall's published terms place responsibility on customers to designate systems they own or are authorized to test and configure the scope before an engagement. The document also says a test cannot certify that an environment is secure or guarantee detection of every vulnerability. Those boundaries matter when autonomous testing becomes an ongoing activity inside an operating company.

The purchasing conversation therefore needs to include the scope, the evidence a team will receive, and how fixes will be revisited as the environment changes. Buyers can evaluate whether the service fits their existing remediation process and who will handle the findings it produces. Those are operating questions arising from the product model, rather than claims that any particular deployment has already solved them.

For CodeWall, the work continues through that second investigation. The customer-success manager helps a team act on the finding, an engineer changes the application, and the authorized attacker returns to examine the route again. The value of the relationship develops as those people learn which conclusions still hold in the software the business is running now.