Torq Acquires Jit: Unleashing the AI SOC Context Graph

Added: 2026-05-25

[00:00:00]Breaking news. Torque has acquired JIT, unleashing the first ever enterprise AI soft, Context Craft, and rewriting the future of security operations. I am here today with David Melamid, co-founder and CTO of JIT, along with Leonid Belin, co-founder and CTO of Torque. Gentlemen, thank you so much for joining us today.

[00:00:19]How are you doing?

[00:00:20]>> Great. Hasn't been more excited for a long time. I want to start off. Tell us a little bit about JIT and what made JIT and Torque the right combination here.

[00:00:29]>> When we ended up talking with Torque, it actually clicked right away. Torque has spent years building exactly what we didn't have. Deep integrations across the security stack and a hyper automation engine that executes cross triage, investigation, and response.

[00:00:45]Apart, we're to great products, but together with the first agentic sock that actually thinks and acts on the same context. not only about the product, it's about the team and about how the team thinks about delivering the value to the customers. What is the place of artificial intelligence in security operations? We've seen a lot of not only synergies but even exactly the same philosophy of what role should it play, how it should be deployed and this fundamental cultural fit is I think what drives this union. The distinction between knowledge graphs and context graphs isn't new, but wiring that into agentic reasoning for secops is what makes this so different. Knowledge graph is basically tells you what things are and how they relate. That's not enough.

[00:01:31]A context graph carries all of that.

[00:01:34]Plus, it adds what an agent actually needs to make a decision, when something is true, where it came from, what the agent is allowed to do, and finally why the agent actually decided what it decided. Like a knowledge graph, you will see identical entities. The context graph will see two completely different stories. One is excfiltration. The other is just another regular Friday.

[00:01:56]>> How does this change what torque agents actually do at triage, investigation, and response?

[00:02:02]>> Malicious actors are now harnessing AI to perform attacks that are way faster, way more complicated, way more multifaceted than ever before. attacks that only like two or three years ago would only be even theoretically accessible to, you know, wellunded state actors are now at a disposal of what we call in the industry script kitties. And what this demands from the defenders is to be able to execute containment and preventative measures. What will ever give you the confidence to do that in a business environment where incorrectly deploying this measure could have a business impact? And the answer is context. This thing actually follows through the whole life cycle of security operations from indeed triage which is an immediate pulse detection and gives us much better ability to provide reliable verdict and then when we do a deeper investigation and understand some root causes of whatever is happening and the end game has to be raising the confidence in ability to deploy protective measures automatically. So guys, what does this mean for CISOs that are using torque?

[00:03:16]>> I would say that the the hardest thing in a sock is not to capture the data.

[00:03:20]Data is everywhere. But I think that the hardest thing is actually to capture and to model the judgment. So on day one to can be calibrated with industry baselines, all your integration, your SOPs is already super useful. The more you use the platform, thanks to the context and the fact that actually the agents are self-improving, Torque will be able to know every override, every approved exception, every deviation from the runbooks and the reason why it was approved and everything lives in the context graph. Context graph is the one thing that guarantees that the grounding of your agentic decisions will not be based on something that is not up to date, obsolete, used to be a good idea, but is no longer such, but is grounded in truth as it is for this very moment in your organizational journey.

[00:04:13]>> What does trust in AI actually look like in secops and why does the torque context graph get us there? Let's start with this. Trust in AI is not a model problem. It's like mostly a grounding problem. You don't trust like a brilliant analyst who who can tell you like why they made a specific call. You actually trust the one who can show you the work. So that's what the context graph actually provide you with. For every decision, you can see exactly what the agent knew, when he knew it, and why it acted. Models will keep getting better. But none of this will matter if the agent is reasoning over the wrong picture of your environment. That's why the context graph matters so much. It's the foundation upon which every future model improvement is actually useful in your sock. I >> I don't take the word trust lightly. I do maybe even take it harsher than most of the people. I'd say that at the end of the day, it's the inevitability of the need to have technology that is not driven by humans deploying in intrusive security preventative measures. We are human beings. The way we raise confidence, especially the technologist among us, is by seeing and understanding how the technology acts. I don't trust non-transparent boxes. I don't trust snake oil. The absolute opposite of that is what we are delivering here with artificial intelligence.

[00:05:35]>> David Leaned, thank you so much for your time. We appreciate you joining us here today and we'll see you next time.