The hardest part of automated security has rarely been the technology. It has been trust. Security leaders have long understood that machines can react faster than people, but many have stopped short of letting software take action on its own, concerned that a false positive or an overly broad response could hurt the business more than the attack.
Upwind, the Cloud and AI security platform, is addressing that concern with Upwind Response, announced on November 12, 2026. Response stops AI-driven attacks within seconds of detection and prevents attackers from moving deeper into a customer’s environment. Crucially, it does so within boundaries that security teams define.
When Upwind sees an attacker try to exploit a cloud application, Response cuts off that attacker’s access to the targeted service and stops the kill chain without waiting for a person to act. The question for security leaders is how much of that autonomy to grant, and Response leaves that decision with them.
The Hesitation Around Automation
Automatic remediation is not a new idea. It has been considered for years as a way to keep pace with attackers. Yet many security teams have resisted it because of two specific fears: false positives, where legitimate activity is treated as malicious, and over-reaching responses, where the fix is far larger than the problem.
Either outcome can create ongoing problems. A blocked customer, a downed service, or a broken workflow erodes confidence in the tool, and once that confidence is gone, teams tend to turn automation off.
Upwind built Response with those concerns at the center.
In practice, that meant three design commitments: narrow actions, full visibility into what was done, and a reliable way to undo it.
Choosing What Runs on Its Own
With Response, security teams decide which threats are mitigated automatically and which require their approval. That choice is not fixed. Teams can start by requiring approvals before Response acts, then allow the service to act on its own as they gain confidence.
Every action is also temporary, visible, and can be easily rolled back. That combination gives teams three forms of control: they set the policy in advance, they can see exactly what was done, and they can undo it if needed.
Proportionate by Design
Response is structured to take the least disruptive action that will stop the attack. It unifies signals across cloud, runtime, code, and identity to understand the full picture of a threat before acting.
When it spots an attacker carrying out a malicious activity, such as attempting to exploit a vulnerability or take remote control of a server, Response blocks the attacker’s access point from reaching the specific service. The block works at the cloud network level and takes effect in seconds, and normal customer traffic keeps flowing.
If the attacker gets further, Response can terminate a specific malicious process and keep it from starting again, avoiding large disruptions to regular operations. It can limit what a compromised machine can reach by trimming its cloud identity so it loses access to object storage and databases. Only as a last resort does Response isolate the machine completely, taking it off the internet and the rest of the network while keeping a secure tunnel into it.
That progression is what separates targeted automation from blunt automation. Each step is narrow, and each step is reversible.
Why Speed Now Matters
The need for faster action is driven by a clear change in attacker behavior. Nearly one in four vulnerabilities attacked in the first half of 2026 was exploited on or before the CVE was published. AI agents can now probe an environment from hundreds of angles, chain dozens of small weaknesses into a breach, and work around the clock to achieve success.
JPMorgan Chase described the situation in a recent post on AI and security: “vulnerability discovery and attack automation are moving at warp speed. Remediation is not.”
Finding risks and detecting attacks remain the foundation of cloud security. But when a new flaw can be used within hours, a scan-and-patch cycle cannot close the window quickly enough. Backlogs of risky settings and unused permissions keep growing at the same time, each one a possible way in.
For security leaders, the question is no longer whether automation is needed. It is whether automation can be trusted. Response is Upwind’s answer to that second question.
Automation that a team cannot see or reverse will rarely be trusted with production systems. Automation that is narrow, visible, and reversible has a far better chance of staying switched on.
The Founder’s Case
Amiram Shachar, Co-Founder and CEO of Upwind, argued that precise visibility is what makes automatic action safe.
“Finding risk will always matter, but against attackers that move at machine speed, defense has to be active,” said Shachar. “Because Upwind sees what is actually happening on the servers running our customers’ applications, we know exactly which address sent the attack and which service it hit. That precision is what makes automatic blocking safe. Every blocked attempt costs the attacker time and money, while our customers’ business keeps running.”
His point speaks directly to the governance concern. Automation becomes acceptable when it is narrow enough that the business keeps running and accurate enough that the right target is hit.
It also reframes the risk calculation. When each action is limited to a single attacker and a single service, the potential cost of a mistake shrinks, and the case for letting Response act on its own grows stronger.
Beyond a Single Organization
Response also contributes to resilience across the wider industry. Each attack is cataloged, anonymized, and added to the knowledge of the Upwind Platform. This allows every Upwind customer to achieve better security through the collective experience of their peers.
Availability
Upwind Response is part of the Upwind Cloud & AI Security Platform and builds on the detections customers already use. Admins can enable it today within the console. Because it uses cloud- and AI-native primitives, there is no extra firewall to deploy.
Upwind Response is available today for all Upwind customers. To learn more or request a demo, visit https://www.upwind.io/response.


