AI Pentesting Aims to Expand Testing Without Increasing Budgets
Summary
The article argues AI pentesting can enable more frequent testing without larger budgets. It outlines risk-based, change-aware testing and describes Aikido’s product and pricing model; cost and speed comparisons are vendor claims.
Key points
- Aikido’s 2026 State of AI in Pentesting report says 76% of surveyed organizations deploy significant production changes weekly or faster, while 21% validate security on every release.
- The article says 42% of teams delay retests because of cost or scheduling, and 40% retest promptly after fixes.
- Aikido says its AI agents map application attack surfaces and validate potential vulnerabilities through exploitation; it claims results can arrive in hours versus a 2.6-week average for manual tests.
- Aikido’s internal benchmarking claims a full-scope AI pentest costs less than half as much as human pentesting; the article notes results vary by location and application.
- The proposed approach assigns testing depth and frequency by application criticality and rate of change, with retests after fixes.
- For continuous testing, the article proposes triggering tests only for security-relevant changes rather than every commit; it says Aikido Infinite offers Git-activity-based estimates, fixed credit prices, and spending controls.
- The article says specialist-led assessments remain important for complex systems and high-assurance requirements.
Article Details
- Topic
- AI pentesting and predictable pricing for more frequent, risk-based security testing
People
James BerthotyFounder and Analyst at Latio, quoted on AI making continuous pentesting more accessible to teams.Katie NortonSenior Research Manager at IDC, quoted on the need to align security testing more closely with software release cycles.Yolanda AmorimApplication Security Manager at Believe, quoted on the difficulty of scaling manual pentesting across multiple regions and teams.