Name: AI Trust and the Rise of Agentic DevOps
Uploaded: 2025-11-18T18:41:03.771Z
Duration: 25 min 8 s
Description: AI Trust and the Rise of Agentic DevOps

Transcript for "AI Trust and the Rise of Agentic DevOps": Welcome to today's webinar on AI trust and the rise of AgenTic DevOps. My name is Drew Pilon, a senior product marketing manager at Cloudbees based in Raleigh, North Carolina, where for the past three and a half years, I have worked alongside our go to market teams to demonstrate how enterprises can gain control of their software delivery efforts through progressive delivery and now AI. I am excited to be joined today by my partner in crime, Jason, who I will now let introduce himself. Thank you, Drew. Hi. I'm Jason Burt. I'm the AI product lead here at CloudViz. I've worked with companies from start ups to enterprises, helping them accelerate their software development with agentic tools, all while building best in class software that's enterprise grade. We're excited to share some of our insights and best practices with you on this webinar. Thank you, Jason. Our focus for today's session is to discuss the current state of AI and software delivery and the role CloudVees plays in it, specifically examining how we enable enterprises to gain control, establish trust, and achieve widespread team adoption for AI. We will then dive into our approach to designing the next wave of AI products and showcase how we're addressing some of these with our new Agenstic capabilities and conclude the webinar with a special invitation. We will reserve time at the end for a q and a, so please include your questions in the chat throughout the session. Let's get started. You can't open your inbox or turn on the news without the mention of AI. And on the surface, individuals are achieving some amazing things. Internally, at Cloudbees, we have hosted our own hackathons and had a dedicated AI week. And the output from these events were eye opening in terms of the efficiency gained. And it forces you to reconsider all of your roles. How do you leverage AI as a companion to help you make you the best at your function? And the concept of Vibe coding is gaining traction. As a product marketing manager with no coding experience, this unlocks perceived superpowers. But with great power comes great responsibility, and this is where the rubber meets the road. When it comes to operationalizing AI in production, projects are still getting stuck in experimental mode. MIT stated that 95% of AI projects are failing to deliver measurable ROI, and Gartner found that 42% of enterprises abandoned their AI initiatives in 2025. So where is the disconnect? We don't view it strictly as a technology issue, but rather as a part of the broader enablement of AI. Before rushing to apply AI to everything, we must first ask ourselves what we're trying to achieve through AI. What having Agenstic systems will enable us to do more efficiently, and how do we ensure that agents are acting as they intended to? Then how will we reconstruct our teams to drive these outcomes? Let's start by examining where trust begins to splinter. From an individual's perspective, AI coding tools promise speed but deliver unpredictability, and that's the disconnect. Developers are wired for determinism, same input, same output, whereas AI operates on probability. And when the results vary, trust breaks and adoption stalls. If developers don't trust the output, they'll avoid it. Worse, they'll spend cycles revalidating, which slows everything down. And the bigger issue is that most organizations are only solving for part of the problem. They add tools without shifting the process or enabling people. Here's what's missing. From a tooling perspective, a secure integrated self serve AI platform, not just another shadow plug in. From a team perspective, the ability to upscale and empower cross functional AI champions and not just throwing tech over the walls. And from a process perspective, pilot fast and the ability to measure real outcomes and scale with guardrails, because speed without safety is risk. Until trust is built across tools, teams, and processes, AI stays stuck in the hype zone, not the delivery pipeline. When you zoom out at the organizational level, there's an execution gap between how AI is enabled at start ups and enterprises. Start ups were born with AI in their DNA. They don't just adopt new tools. They design around them. Agenstic's Suisse aren't sidekicks, but they're central to the workflow. This is evident in a recent survey that found that two thirds of AI native start ups have AI write most of their code. Meanwhile, enterprises are still trying to wedge these agents into delivery pipelines built for and by humans and are doing so in solid approaches when dealing with governance, risk management, and other compliance requirements. The result is friction, shadow tools, and wasted potential. They have to make AI work within their existing heterogeneous tooling mixes, and they don't have time to completely rip and replace to start over. So how can enterprises leverage AI to close this gap while tackling the age old challenge of speed versus safety and flexibility versus control? CloudBeasts has defined our vision and mission statements to align here. With over fifteen years of DevOps expertise, we have witnessed multiple market disruptions, and the results are clear. Forcing an all in one approach simply does not work. You must leave a choice and ensure that proper guardrails and telemetry are in place. At CloudViz, we are enabling enterprises to innovate faster with confidence and control. We are helping to turn AI from an experiment into an enterprise operating system for software delivery, and this acknowledges the technologies will continue to come in and disrupt, but we have a long term view for how these things will coexist. CloudViz Unify is our answer to how this gets delivered. It sits above your delivery tools to unify security and delivery telemetry into a single normalized data model. This provides governance and delivery awareness without touching the underlying pipelines. It creates a coordinated system of record that aligns your existing investments without forcing standardization or migration. And this is critical because a recent Cloud based commission survey showed that 92% of enterprises report achieving greater delivery efficiency by integrating tools rather than replacing them. I will repeat that. 92% of enterprises report achieving greater delivery efficiency by integrating tools rather than replacing them. So bring your build tools like Jenkins or GitHub Actions, your SCMs, your scanners, and your agents. CloudVees Unify collects the data from all of them and lets your team or agents orchestrate them from one place. It's not just about integrating tools anymore. It is about intelligently connecting all of your pipelines into the ultimate dataset that feeds your teams and AI agents, empowering them to turn fragmented delivery into coordinated, secure, and traceable flow. This transforms developer toil into strategic acceleration, which leads us back to the trust gap. We define trust as systems that act responsibly, explain their decisions, and stay aligned with policy and purpose. Trust is achieved when you combine control plus tailored context. Let's break these concepts down further. The control plane does two things that every enterprise needs. It controls the chaos caused by tools for all and unlocks the full value of your delivery stack without forcing costly migrations. This sets the foundation for the context plane to provide agents the understanding and the intelligence they need to act safely. Through tailored context, CloudVees ensures that agents have tight windows to accurately collect information while establishing the guardrails for responsible agentic AI across all levels. With trusted AI as your guide, let's discuss how you can drive change and adoption across all organizational levels to create shared context. Understand that approaches taken must differ based on the level. The org drives change and individuals are gaining hands on experience, but Teams is where the real adoption happens through the sharing of best practices and enablement strategies. Let's make this real. Imagine that you're a VP of software engineering looking for best ways to introduce a to models into your daily workflows. You need to decide what task to offload to agents and how these will interact with your human developers. To help align expectations, start by treating agents as junior developers joining your team. With this construct in mind, what task are you comfortable sharing with them? These tasks are often well defined, role based, and broken into smaller steps. By offloading these tasks, what does that unlock for your more senior team members? And how does this impact your plans for team dynamics when you can have seemingly an unlimited amount of junior developers? Build out a thirty, sixty, 90 plan for onboarding your agents to ensure you don't get lost in the hype of Agenstic AI and that you are able to align expectations with stakeholders when defining projects to meet your stated ROI goals. With this crawl, walk, run approach in place, you equip your teams for actionable learnings and a strong Agintiq foundation. And with this all in mind, I will now turn it over to Jason who will cover how CloudVisa is approaching this from a product perspective. Thank you, Drew. We see software development as a team sport. Today, to produce enterprise grade software, you need platform engineers to provide stable, repeatable, and reliable systems. You want QA engineers who can help verify if the software is consistently conforming to a customer standards. You want amazing security engineers to help keep track of changing regulatory standards, company controls, CVEs, and other security concerns. And for complex systems, you want release engineers to help keep the train moving to reduce the friction. This context of worldview organization and application is too much for one individual or team without this augmentation. So what is the future of software development? With AI, we have the opportunity to enable developers with best in class team members, enabling these teams to operate seamlessly together. First off, AI first responder acts as a context orchestrator between teams' triaging requests. Reducing time for remediation is really reducing time to context. We see agentic QA as a tool that can take in requirements and work seamlessly between the team and the organization, building best in class software for your customers. We see AI security assistance going beyond triage and really understanding the SLAs and remediation requirements at the right stage for your teams, tailoring the right security controls for the right stage of the release process. Finally, on the CIC and release side, we see the ability to enable agentic suite and developers to go from prototype to production seamlessly, keeping your application in check the whole time. To harness these new capabilities while ensuring trustworthiness, it's not enough to just have an agent. You have to add AI on top of your team's workflows, meaning that you can help maintain that mental model across the systems, teams, tools, and agents. CTOs we talk to talk about technical debt, but there's also context debt. This context debt adds more overhead to your teams and blocks the ability to reinvest your time. So to unlock that time, we're adding an AI planner, which provides the ability to build scenarios to enable your team to adopt these tools with AI first responder. This reduces the comp complexity of managing this mental model. The tension of system, team, and market being out of alignment is oftentimes what can cause these teams to shift focus. Also, with AI guardrails, we wanna protect your entire context plan, making sure all your agents are working properly, skilled properly, and do do not have misguided definitions. If it's a tire developer who's a new parent who's pushing some code on a Friday or an external threat actor trying to do prompt injection in your downstream documents, we wanna make sure that AI guardrails are making sure that your team is set up for best in class success, monitoring and maintaining that context plane, and setting up these mission critical teams so that they can do their best work, every second of the day. As DevOps reshapes software development, Agentic tools has reshaped the industry with tailored Agentic partners. We wanna provide you with the tailored right tool for the right time. This requires planning at a team and order level going beyond simple insights. With scenario planning, you can come up with a plan to roll out things like Agenstic Security across your entire organization. In complex markets, plans always change, so AI first responder will help your teams adapt and adopt to whatever comes up. And finally, with Unify's MCP server, we provide that single connection point for your developers so you can use their preferred tool to tap into the power of Unify's AI abilities throughout their entire journey. So with that, let's dive into a few ways that we are updating our product experience to help drive teams and organizations into the future. This demo shows you how AI first responder with Unifi MCP server enables your team to rapidly go from prototype to enterprise grade production. As we mentioned, agentic tools are changing the way we code. You can use your favorite agentic tool of choice such as AWS Q, Cloud Code, Gemini. You name it. You can use any of these tools with Unify and Unify AI. As an example, I copied over my demo script into my IDE. From here, I used AWS Q and kicked off a process to basically build my own custom application. This is basically TDD or spectrum in development. This process took me about thirty minutes while I was making coffee. As you can see, once I was done with this stage, I pushed my repo to GitHub. From here, I used Unify's MCP server to connect the project to my other teams and tools and also to set up enterprise grade controls so my other team members who join can contribute as well. Here, you can see I added Jack, who's one of our engineering leads and UX leads on the project. In addition to this, as Unify helps me initialize my projects and connect the dots between my teams and tools, it also set up a Slack channel for this AI first responder demo. As a new project, AIFR will make sure that trusted context is being built between your domain experts and your knowledgeable team members. So if you have technical PMs, tech writers, engineers, whoever's working on the project, they can help contribute their knowledge to this group. So on the right, you can see I have a number of markdown files that exist. I use Unify's MCP service to map my existing repository and push relative files to AFIR so they can be indexed and used both by team members and agent and agent of tools. Here you can see my AI demo app script. So I'll go over to AIFR, and I'll engage ask mode. There's a number of modes that AIFR comes in to enable your team members so you don't have too many messages overloading your teams. So moving into ask mode, I'm going to ask you about this new service that I launched. Great. So in this mode, AFR will ask you if it wants to move forward or if it can move forward. So I wanted to start to respond to my question. There we go. So as you can see, AFR has responded with the document that I've uploaded that provides the context and the overview of the application that I just created. Now I might be away. I might be on a customer call. I might be sick. So you can set AIFR for auto remediation. This allows other team members who might not be technical or have access to the repository or other technical documentation to ask questions on it. As an example, Drew, the PMM who is on this call, we partner all the time on writing copy for services. So Drew can ask AIFR for information about the service without pinging the technical team or the PM. Great. So here you can see AIFR has automatically responded to his questions. While you can pin the service at an application level, you can also uplevel it so it can provide information about your organization, your division, or other information available. So here, Drew asked about Unify, and AIFR is just one of the many features that we have in Unify. Once there's trust in place, AIFR can be used to actually go and update the documents similar to GitOps. It also can be used for auto remediation and triage of other issues, but we wanted to start with that safe knowledge base to get started. Today, we are showing you the knowledge triage aspect of the service. Once enabled, AIFR can actually go beyond that, actually providing direct auto remediation. If it's updating your Git repository, it can create a pull request for you to review. It can also be wired into Unify to automatically fix pipelines and other issues in your DevOps process. All of this collapses the feedback cycle, greatly enabling your teams to plan together and work in parallel. We have more coming from the team, and we're excited to share more of this agentic enablement journey with you. And with that, we'd like to introduce you to the CloudVees AI design partner program. Join us in driving trusted AI adoption, both in your company and in the industry. This program gives customers access to and access an early influence to our AI capabilities before release, helping influence our roadmap. We also wanna co create standards and best in class software practices that we can share internally with your company as you drive change in your organization, but also share these practices externally with the rest of the, the world and the ecosystem. And finally, we know that time is precious. It's one of our most precious resources. And so we wanna make sure that we're engaged with you. We're a three month period, rapidly providing feedback where you can up level your organization, but also, rapidly gain through the value of of our products and ecosystem. Great. We look forward to you joining the partner program. So you can either reach out directly to CloudVees members that you know today, like the sales teams or other groups. You can also sign up at ai.cloudvees.com. Thank you, Jason, for sharing your perspective and for going through the AI first responder demo and also the invitation to join our AI design partnership program, which sounds like a great opportunity to collaborate with your team. I will now bring this home with the conclusion slides. For over fifteen years, CloudVease has been a trusted partner in the DevOps space, empowering thousands of global developers to build, test, and deploy the world's most mission critical software faster, safer, and at scale. With over $150,000,000 in ARR and profitable, we prove that growth and discipline can coexist. We are backed by a thriving developer community, and we are redefining the future of DevOps, committed to a world where AI and automation unlock developer creativity, not to constrain it. And we're trusted by the global 2,000 across all industries. We are excited to continue the conversation with you on the future of Agintiq DevOps and how CloudVees can help you win. To reinforce a few key takeaways from today's session, Agintiq AI is still in the early stages for software delivery, with enterprises often getting stuck in the experimental phase. Breaking through requires organizations to be intentional in how they enable AI across teams to establish trust and grow adoption. Cloud b unified is our answer to establishing trust for enterprises with the unique combination of control plus tailored context. And we have launched the limited preview for AI first responder and are looking for design partners to help build the future of agentic DevOps with unparalleled access to our product and engineering teams. Apply to join as an AI design partner today. Thank you for joining today's session.