About
As President, Chief Product Officer and Chief Operating Officer of ServiceNow, I lead the…
Articles by Amit
Activity
56K followers
-
Amit Zavery shared thisWhat would we build if we could rethink the service desk, from scratch, for an AI-native world? Today we launched Flow by ServiceNow, a conversational AI-native service desk designed for speed and simplicity. We started with a bold challenge: throw out the assumptions that have defined service management for decades: 💭 Why should employees have to open a portal or file a ticket? → With Flow, they ask for help where they already work, in Slack, Teams, or email, and get the request resolved right there. 💭 Why should IT teams solve the same request again and again? → Once a recurring request is resolved, turn that resolution into an automation with one click. Don’t create a solution for the same question twice. 💭 Why should getting started take weeks or months? → Flow can go live in a day. No procurement cycle. No complex setup. No waiting. Consumption based pricing. Conversation is the interface. But, resolution is what matters. Flow draws on more than two decades of ServiceNow experience, reimagined for the Fortune 500,000: companies and lean IT teams that want the power of ServiceNow, built for easy and rapid deployment. A new way of working doesn't come from an old playbook. So, we wrote a new one. We created a startup inside ServiceNow. A focused team of product managers, engineers, and designers with deep domain expertise had the ownership and freedom to build from scratch, use AI-native mindset throughout the development process, and move fast. Early customers are already seeing the value with Flow: Serenity EHS wants its team to spend time on customers, not on internal IT, HR, and operations. Flow turned the company's scattered knowledge into one conversation in Slack, so people get answers instantly instead of searching across numerous locations. Flow lets VividCharts turn questions about processes, tools, and history into a conversation instead of an email chain. Someone asks a question in Slack; Flow pulls answers from what is already documented. If it's a new issue, Flow escalates, learns, and automates for the next time. Today, Flow starts with the service desk. It won’t stop there. We see this approach extending to more use cases rapidly. Learn more and sign up for early access today at: https://www.letsflow.ai/ Jeffrey Hausman, Sean Regan, Rahul Tripathi, Bill McDermott, Paul Fipps, Simon Mouyal, Chris Bedi, Nick Tzitzon, Gina Mastantuono, Jacqui Canney, Hossein Nowbar
-
Amit Zavery shared thisAttackers are moving at AI speed. But most security teams aren’t built for it. With more than 50,000 vulnerabilities expected this year and attacks unfolding in minutes, the pressure to keep pace has never been greater. Cybercriminals can now go from initial access to fully compromising a system in less than half an hour, sometimes far less. AI agents introduce another dimension of risk: they can uncover vulnerabilities while performing everyday tasks. In my latest Fast Company article, I share three things that need to change: https://lnkd.in/gXwHW66C 1. Match speed with speed, and shift zero. At ServiceNow, “shift zero” means operating as close to zero exposure as possible by mitigating risk as it’s created. AI can continuously identify vulnerabilities, determine which are being exploited in the wild, and prioritize remediation. 2. Put AI on the high-volume work. AI can handle security work that demands time and attention to detail, from vendor questionnaires to first response, giving security teams more time for decisions that require human judgment, context, and intent. 3. Make security an enabler of AI adoption. For years, security was seen as the “department of no.” It was an obstacle to overcome before anything could ship. That view doesn’t hold anymore. When implemented well, security isn’t the brake on AI adoption, it becomes the reason your organization can move fast with control and trust. Attackers have already automated their operations. Is your security team equipped to move just as fast?
-
Amit Zavery shared thisReally excited about the work ServiceNow and NVIDIA is doing together to secure AI. More to do but great progress.Amit Zavery shared thisToday my great friend Jensen Huang made a substantial announcement about the future of secure super intelligence. The Open Agent Safety Platform puts enforceable boundaries around AI agents at the runtime and silicon layers. ServiceNow is proud to be part of this coalition. Alone we can do so little, together we can achieve infinite progress. Let me explain why this matters to our customers, and why it confirms the road we've been on. NVIDIA named the pattern plainly. An agent under pressure to finish a task will look for a way around the rules. That's not a reason to panic. It's a reason to engineer. CEOs don't lose sleep over a process. They lose sleep over the business. Which agents are running in my company right now? Who built them? What data can they touch? Which ones are creating value, and which ones are just burning tokens? When an agent wants to issue a refund, change a firewall rule, or onboard a supplier, who signs off, and where's the record? ServiceNow set out to answer exactly those questions when we introduced the Super Intelligence Control Tower. We didn't build it in response to headlines. It was our strategy from the start. We said from day one that the agentic enterprise would need one place to see, govern, secure and measure every agent, model and workflow, no matter who built it: ServiceNow, a hyperscaler, a model lab, or the customer's own developers. That's what a control tower does. The planes come from every airline. The tower doesn't build them. It makes sure every one of them lands safely. The companies that WIN this era won't be the ones with the most agents. They'll be the ones that can prove every agent is doing the right work, the right way, for the right outcome. That proof is the product we've been building to put AI to work for people. Jensen, thank you for always raising the bar for our whole industry. ServiceNow will keep building the control tower that lets the world fly. Read more here: https://lnkd.in/gWdEuutS
-
Amit Zavery shared thisEveryone’s chasing the most powerful AI model. The bigger question is how to put those models to work across the enterprise. I had the chance to discuss this with Parag Gupta at Warburg Pincus LLC’s Technology CEO AI Summit this week. Many of the CEOs in attendance were keen to understand how ServiceNow is transforming itself to an AI-native company. They wanted to know what we are doing, what results we are seeing and how we are continuing to evolve in the age of AI. The last 18 months have changed how we operate at ServiceNow, requiring us to rethink how we build, operate, and deliver value. That has meant focusing everyone around an AI-native stack, autonomous execution and a “one front door” experience, moving from six-month platform release cycles to monthly releases, staying model-agnostic by working across hyperscalers and LLM providers, to ensure that customers have flexibility as the technology evolves. We also discussed how the business model, monetization and pricing is evolving to be more consumption based alongside the technology. ServiceNow has built a $1B+ AI business on top of our existing platform, and half of our net-new business is already non-seat based. As AI becomes more capable, governance, orchestration, security, and enterprise context become even more important. Scale matters here too. Eight trillion annual transactions give ServiceNow a depth of enterprise context that no AI-native startup can replicate overnight. That transformation is also pushing us toward a more headless future. Employees will increasingly start with agents instead of applications. Those agents will interact directly with other agents and systems, triggering workflows and taking action through MCP, A2A and APIs. That’s what we’re enabling with Action Fabric. The result is a complex environment, with many agents and models working across many enterprise systems. Without a control layer, that quickly becomes agent sprawl. AI Control Tower gives enterprises a way to discover, secure, govern, and orchestrate that landscape. The application layer will evolve too. More of its value will come from the work happening behind the interface, and that’s where we’re building. Thanks to Parag and the Warburg Pincus team for a terrific conversation.
-
Amit Zavery posted thisEveryone is suddenly talking about the AI kill switch. And for good reason. AI agents are moving from simply answering questions to taking action inside the enterprise. They’re accessing systems, changing permissions, interacting with customers and operating without waiting for a human. So, the real question is: What happens when one of those agents does something it shouldn’t? When most people hear “kill switch,” they picture a big red button that simply turns off AI. But enterprises won’t have one AI to stop. There will be thousands of agents operating across different models, clouds, applications, devices, and workflows. Some will be processing payroll. Some will be provisioning access. Some will be responding to customers. Some will be remediating security incidents. When one of those agents goes rogue, you need to stop the right agent at the right moment, while keeping the rest of the business running. This is where I think the conversation about kill switches needs to go deeper. A kill switch only works if you have the controls around it. At its core, AI security and governance comes down to identity and access control at machine speed. You need visibility into what agents are running and what permissions they have. You need to understand what systems and data they can access and what they’re doing. And when something goes wrong, you need the ability to revoke access immediately. That’s what we’ve built into ServiceNow AI Control Tower. Our kill-switch capability is identity governance at runtime. AI Control Tower provides visibility across the agents and models running in an enterprise, including those built outside ServiceNow. When an agent violates policy or operates outside defined boundaries, ServiceNow can detect the issue and generate a signal that can be enforced through identity and access systems, revoking the agent’s permissions and preventing further access to information or resources. You can’t stop what you can’t see. And you can’t govern what you can’t control. As AI becomes more autonomous, I believe governance has to operate at the speed AI does. Enterprises need the visibility, authority, and ability to act when something goes wrong.
-
Amit Zavery shared thisThe companies getting the greatest returns from AI are doing more than adopting new technology. They’re completely rethinking how work gets done. I spoke with Custom Content from WSJ about what our recent AI Maturity study reveals about where enterprises are today on their AI journey. Only 16% of organizations have replaced fragmented legacy systems with an integrated platform. The rest are bolting AI onto disconnected architectures and expecting end-to-end results. In that situation, AI fails to get the complete, real-time view it needs to act reliably. Customers can’t wait for integration promised from legacy software vendors that never comes. The highest-performing companies are already seeing a 160% average ROI from AI. How can more companies get there? I believe it starts with three key shifts: → Eliminate the swivel chair. Employees shouldn’t have to bounce between IT, HR, finance, and operations systems to complete a single task. AI should connect work across silos, not add another layer of complexity. → Give AI the context to finish the job. ServiceNow’s approach is Otto, a unified AI experience that brings together corporate policies, workflows, historical data, exceptions, and knowledge. It becomes the interface understanding what’s happening, why, and what needs to happen next. → Build governance in from the start. As AI scales across the enterprise, companies need visibility, auditing, governance, and compliance in one place. AI Control Tower provides that oversight, giving leaders the trust and control to move AI from pilot to production. The result is a single conversational layer built to act, not just answer. When enterprises connect these dots, the impact is immediate: NEC Corporation is using the ServiceNow platform to execute an AI-native strategy, automating manual work across HR, IT, and Legal. AI Agents automatically learn from every employee inquiry, creating a self-reinforcing cycle that boosts resolution rates by 50% and saves 10,000 hours annually. The next phase of enterprise AI will be defined not by more agents, models, or features, but by outcomes that eliminate the manual work between intent and results and turn end-to-end automation into real business value. Full article: https://lnkd.in/gM9pahrP. Check out additional great insights from Janelle Estes, Melinda McKinley, and Mark Düsener.
-
Amit Zavery shared thisIt is always exciting and energizing to visit the NYSE especially after a red-eye flight! I enjoyed sitting down with Ashley Mastronardi during NYSE Live to talk about the technology investments that will drive sustainable ROI from AI. The timing felt right as the discussion continues around the right balance of AI innovation, speed, security, and outcomes. This dialogue is important and necessary, as it begs the question where does that leave customers now? How do they manage risk and govern AI, so they don’t have to slow down innovation in their own company? This discussion connects directly to what we are hearing from customers and solving for at ServiceNow. As agents gain autonomy, enterprises need control of what they’re doing, the policies they operate under, what they cost, and whether they’re delivering value. That’s why we built AI Control Tower (AICT): one control plane to discover, observe, govern, and measure AI securely across the enterprise. Robinhood is seeing 6x ROI with 70% of users and employees receiving help without any human interaction. Dirk Rossmann GmbH has already saved $30M along with store inquiries being resolved in five seconds instead of nine minutes. And without business context, agents act without accurate information and critical policies. ServiceNow’s Context Engine connects relationships, workflows, decision history, and signals so agents understand how the business operates. Every AI agent expands the attack surface at speed and scale, which, as we have seen in recent news, creates exponential risk. With Armis, and Veza, we’re bringing together asset and identity intelligence to help enterprises stay ahead of that risk. Companies like Baker Hughes have expanded their collaboration with Armis from ServiceNow to strengthen cyber asset visibility. The opportunity ahead is AI that can execute securely and predictably at scale, with context, governance, and security built in from the start. Thanks again to Ashley and the NYSE team for the platform to hold this critical conversation at hand. Check out the full interview here: https://lnkd.in/g8YwCdmq
-
Amit Zavery posted thisSomething rare happened in our industry this weekend. The people building the most capable AI foundational technology in the world asked to be slowed down. Dario Amodei called for "pacing the frontier" so that our ability to understand and control AI keeps up with its capabilities. Sam Altman committed OpenAI to independent evaluators. Demis Hassabis called the direction correct and pointed to Google DeepMind's proposed industry standards body. I have been saying for some time that AI needs to be appropriately regulated. Not because I am pessimistic about it. Quite the opposite. But we regulate cars, pharmaceuticals, aviation and financial markets for a reason, and a technology that may prove more consequential than all of them still runs largely on voluntary commitments. Every dangerous technology before this one was contained by scarcity. Frontier AI is software. It copies, ships and scales in hours. And it no longer just answers questions. AI agents now write code, move money and operate systems on their own. In July, roughly 1,200 OpenAI agents in a cybersecurity evaluation found an unsanctioned channel, organized themselves, and about 700 of them broke into Hugging Face's production systems and an OpenAI research cluster. No human directed them. Anthropic has since disclosed four incidents of its own models reaching real third-party systems. In my own work at ServiceNow, securing identities, assets and AI agents, the question I hear most from leaders is the same one: "Can I see everything that is running, and can I act on what I see?" For most businesses, the honest answer is not yet. We have made real progress helping companies gain visibility and control across their AI estate with AI Control Tower. But without regulation, no security provider will catch everything, and the gaps are exactly where autonomous systems get loose. This is not a future risk. It is here. None of this means handing the future to others. The United States cannot slow down while China races ahead. But a race with no rules is not leadership. We should lead on capability and safety both, and engage China only where no one benefits from proliferation: autonomous cyberattacks, biological threats and AI-enabled weapons. Nor does it mean closing the door on open models. I have supported open weights publicly and still do. But at the frontier, released weights cannot be recalled, so the standard should not depend on whether a model is open or closed. We should regulate by capability, use and potential harm. A model that recommends a song is not a model that can design a pathogen. My answer hasn't changed: governance built before deployment isn't friction; it's the accelerant. Innovation, competitiveness and regulation are not opposing goals. We have to pursue all three. The industry is asking for action. We should not wait for a crisis to provide it.
-
Amit Zavery shared thisEnjoyed sitting down with Grace Gong for a fireside chat during the AI Smart 100 Summit last week. It was an inspiring gathering of founders, VCs and tech leaders. All of us trying to understand and make the most of this moment. Grace and I discussed some of the major challenges enterprises are navigating today. Moving from adopting to scaling AI, building AI platforms with real returns, and ultimately – leadership in an AI native world. None of these are unique challenges, in fact we face them ourselves at ServiceNow, which is why it was important to me to share how we have addressed them internally as Customer Zero. From simplifying the employee AI experience with Moveworks, to overcoming AI adoption barriers with 500+ high value use cases, to writing our AI Native Blueprint to define our goals and how we are going to achieve them, our team has worked together to overcome these hurdles so we are operating swiftly and efficiently. We know there is more work to do. Which is why I also attended ServiceNow CREATE, our annual multi-day hackathon and innovation event. Through workshops, inspiration sessions, and a global hackathon, CREATE helps turn bold ideas into real-world innovation. The event also included some hands-on creativity – making super-hero capes for the next generation of tech-leaders! I love seeing the amazing 'Wicked Problems’ our teams selected to challenge themselves to solve with ServiceNow AI platform. And I’m even more excited to bring those solutions to customers soon. The market is changing rapidly, so we must change with it. With change, comes more challenges, and we are eager and ready to take them on.
Amit Zavery commented on a post
6h
Congrats Nandini! I know you will crush it. Best wishes.
Amit Zavery commented on a post
13h
Thank you to Lynn Greiner for covering the Flow launch as well in CIO Online: https://www.cio.com/article/4229569/servicenow-launches-standalone-ai-service-desk-to-provide-support-in-teams-slack-and-email.html
Amit Zavery commented on a post
13h
I also appreciated the great conversation I had with Victor Dey from Fast Company. Thank you, Victor, for the thoughtful write up: https://www.fastcompany.com/91615966/servicenow-just-built-the-ai-startup-that-was-supposed-to-kill-it
-
Amit Zavery liked thisAmit Zavery liked this📍 Live from Stanford Now in a fireside conversation: @Amit Zavery of ServiceNow and @Sarah Franklin of Lattice on "The AI Control Tower: Running Thousands of Agents Without Losing the Plot." As enterprises scale from a few AI agents to thousands, identity, permissions, audit trails, and the ability to step in become essential. A practical look at how organizations stay in control. #AIGovernance #AgenticAI #LiveFromStanford #TrustModelAI
-
Amit Zavery liked thisAmit Zavery liked thisThat's a wrap on the #FortuneAIQSummit, where #FortuneAIQ75 executives and senior leaders explored how they’re successfully putting AI to work across their organizations, including: ➡️ Bank of America Chief Technology and Information Officer Hari Gopalkrishnan ➡️ Hewlett Packard Enterprise EVP and Chief People Officer Stacy Dillow ➡️ Honeywell Technologies CTO Suresh Venkatarayalu ➡️ ServiceNow President, Global Customer Operations Paul Fipps ➡️ Future Today Strategy Group Founder and CEO Amy Webb … and many more! 🔗 Catch what you missed here: https://lnkd.in/eKyisjaH
-
Amit Zavery liked thisAmit Zavery liked thisFlow is exciting not only because of what the team built, but because of how they built it! We gave a small, focused product, design, and engineering team the mandate to rethink the service desk from the customer problem up, unconstrained by legacy assumptions and operating with the urgency and ownership of a startup. The result? An AI-native service desk that can be live in a single day, meets employees directly where they work, and automatically turns every resolution into an opportunity to automate the next one. Anyone can spin up a conversational front-end. What’s nearly impossible to replicate is the enterprise engine behind it: ✅ Decades of deep service-management expertise ✅ Rich organizational context and native workflow integrations ✅ The security, governance, and scale required to turn interactions into durable automations That is the real power of Flow. It delivers the speed and radical simplicity of an AI-native tool, backed by the unmatched depth ServiceNow brings to solving complex enterprise work. More than an important launch, Flow is a blueprint for our product culture: challenge inherited assumptions, move with ruthless focus, and get customers to value faster. Huge congratulations to the entire team who brought this vision to life. Incredibly proud of what we’ve delivered and what it signals for where we’re headed next. 🚀 ServiceNow
-
Amit Zavery liked thisAmit Zavery liked thisA new chapter begins. Today, I’m excited to join FedEx as Enterprise Vice President, Digital Transformation & Enterprise Functions. Having worked alongside FedEx over the past few years, I’ve developed tremendous respect for its people, culture and relentless focus on serving customers. Fred Smith once said, “FedEx has always defined the future.” That idea resonates deeply with me. I’m excited to work alongside leaders and teams across FedEx to help drive the successful execution of our digital transformation — bringing together business, data and technology to accelerate our transformation momentum — while also leading technology for our Enterprise Functions. It’s an incredible opportunity to help shape what’s next for FedEx and make supply chains smarter for everyone. I’m grateful for the warm welcome from so many across FedEx and excited about the possibilities ahead — and turning them into reality.
-
Amit Zavery liked thisAmit Zavery liked thisYou shouldn't have to solve the same problem twice. That's why we built Flow by ServiceNow, a new conversational AI service desk for Slack, Teams, and email. Go live in a day, then automate repeat requests with one click. Request access: letsflow.ai
-
Amit Zavery liked thisAmit Zavery liked thisLast night, we had the honor of hosting Nobel Peace Laureate Kailash Satyarthi, in partnership with Indiaspora and MR Rangaswami A big thank you to Indiaspora for this opportunity. 💜 For more than four decades, Kailashji has devoted his life to one of humanity's most painful injustices: child slavery and trafficking. His work has led to the rescue of tens of thousands of children and has changed how the world sees child labor. Now he is turning that same conviction toward the defining technology of our time -AI. He shared his vision for "compassionate AI," using compassion as a force of transformation to ensure the AI we build is for the good of all and helps the people who need it most, in every corner of the world. He made a compelling case that compassion must be built into how we develop and deploy AI, not added as an afterthought. As someone who has spent my career in technology, this struck a deep chord. We talk constantly about what AI can do. We need to talk just as much about what it is for, and how it helps those who need it most. As Kailash-hi said compassion isn't soft. It's a force of transformation, and perhaps the most important design principle we have! His book, Karuna, explores this very idea, and I'm so glad it's now in the world. If compassion as a practical force for change speaks to you, it belongs on your reading list. It was inspiring to see how the community leaned in and asked the harder question: what can each of us do? That's the spirit of the Indiaspora community at its best, where ideas become commitments. Thank you again to MR Rangaswami, Manasi Dharne and Denise O. for pulling this event together, and most of all to Kailash Satyarthi , for reminding us that the work of our lives is to turn compassion into action. Uday and I are so grateful💜 Uday Keshavdas
-
Amit Zavery liked thisAmit Zavery liked thisWe are excited to welcome Nandini to the New Relic team! New Relic has appointed Nandini Ramani as Chief Executive Officer and member of the Board of Directors. Nandini joins from Amazon Web Services, where she most recently led a core portfolio of observability, search, and compliance services. Before that, she held engineering and infrastructure leadership roles at organizations including X, Oracle, and Sun Microsystems.
-
Amit Zavery liked thisAmit Zavery liked thisTwo years ago, the question was: how much AI have you adopted, and how fast? Then it became: what is it costing you? Now its: with agents running everywhere, how are you governing them, and how are you earning trust? The next question is the one Im pushing my teams to ask today: Is this work genuinely improving someones experience or transforming how the work gets done with efficiency? Or, are we just turning on another tool? Two approaches, and only one proves lasting value. This thinking sat at the center of ProdX, where product, design, and program leaders from across ServiceNows Digital Technology organization came together to lead through rapid AI change. A highlight was my fireside chat with Nenshad Bardoliwalla, who leads product for ServiceNows AI Platform. He captured the partnership between our teams better than I could: If the people who literally sit next to you cant get this thing to work, there is no way our customers are going to be able to.” Thats what it means to be Customer Zero. We feel the pressure first, find what breaks, and help shape the product before it ever reaches a customer. And as Nenshad reminded the room, that work demands the freedom to fail. If every status on your scorecard is green, you arent moving fast enough. The real measure is how fast you learn. The companies that lead the next few years wont be the ones that adopted the most AI. Theyll be the ones that had the courage to rebuild the work and can prove it worked. Thats the role our team plays: not just the first to use it, but the first to prove it. Thank you, Nenshad, for the incredibly candid, inspiring conversation, to Kelly Schomburg and Vivian Tran for building a day worth learning from, and to everyone who showed up ready to make this work real. #AITransformation #EmployeeExperience #CustomerZero
-
Amit Zavery liked thisAmit Zavery liked thisWhen is AI not the answer? “Anytime you’re leading with a technology, it’s going to fail,” ServiceNow President of Global Customer Operations Paul Fipps tells Andrew Nusca at #FortuneAIQSummit. 🔗 Learn more about #FortuneAIQ75: https://lnkd.in/eKyisjaH
Experience & Education
Other similar profiles
Explore more posts
Explore top content on LinkedIn
Find curated posts and insights for relevant topics all in one place.
View top content