AI is rewriting cybersecurity faster than enterprises can respond.Our new Kalaari report captures what CISOs are actually struggling with, and where they believe the next wave of cybersecurity companies will be built.

‍
𝐂𝐈𝐒𝐎𝐬 𝐭𝐨𝐥𝐝 𝐮𝐬 𝐭𝐡𝐞 𝐠𝐚𝐩𝐬 𝐚𝐫𝐞 𝐜𝐥𝐞𝐚𝐫: AI systems introduce risks traditional tools can’t see, identities are multiplying beyond control, and SOC teams can’t trust copilots without guardrails. These aren’t theoretical concerns, they’re lived operational pain.

𝐅𝐫𝐨𝐦 𝐭𝐡𝐨𝐬𝐞 𝐜𝐨𝐧𝐯𝐞𝐫𝐬𝐚𝐭𝐢𝐨𝐧𝐬 𝐚𝐧𝐝 𝐫𝐞𝐚𝐥 𝐜𝐚𝐬𝐞 𝐬𝐭𝐮𝐝𝐢𝐞𝐬, 𝐟𝐢𝐯𝐞 𝐨𝐩𝐩𝐨𝐫𝐭𝐮𝐧𝐢𝐭𝐲 𝐚𝐫𝐞𝐚𝐬 𝐬𝐭𝐚𝐧𝐝 𝐨𝐮𝐭:
- 𝐀𝐈 𝐑𝐞𝐝 𝐓𝐞𝐚𝐦𝐢𝐧𝐠: prompt-injection, RAG poisoning, jailbreaks, and model-misuse testing that legacy AppSec cannot cover.

- 𝐀𝐈 𝐒𝐮𝐩𝐩𝐥𝐲 𝐂𝐡𝐚𝐢𝐧 𝐒𝐞𝐜𝐮𝐫𝐢𝐭𝐲: scanning and attestation for models and datasets exposed by incidents like malicious PyTorch/HF artefacts.

- 𝐒𝐡𝐚𝐝𝐨𝐰 𝐈𝐝𝐞𝐧𝐭𝐢𝐭𝐲 𝐆𝐨𝐯𝐞𝐫𝐧𝐚𝐧𝐜𝐞: discovering and controlling AI-spawned tokens, service accounts, and agent credentials.

- 𝐃𝐚𝐭𝐚 𝐏𝐫𝐨𝐭𝐞𝐜𝐭𝐢𝐨𝐧 & 𝐀𝐈 𝐆𝐨𝐯𝐞𝐫𝐧𝐚𝐧𝐜𝐞: provenance, unlearning, cross-border compliance, and auditability for DPDP/AI Act environments.

- 𝐑𝐞𝐥𝐢𝐚𝐛𝐥𝐞 𝐒𝐎𝐂 𝐀𝐮𝐭𝐨𝐦𝐚𝐭𝐢𝐨𝐧: confidence scoring, safety checks, and explainability layers to tame hallucinating copilots.

These are the fault-lines CISOs feel today and the spaces where founders can build the next generation of security infrastructure.