Each CIO and CISO we communicate with describes the identical paradox: AI is now central to their transformation agenda, but the quickest solution to derail that agenda is to lose management of AI. As generative AI, agentic methods and embedded AI options unfold throughout the enterprise, leaders are now not asking in the event that they want AI safety; they’re asking what sort of AI safety technique will really scale.
Gartner® has revealed two current studies that validate this actuality and description the strategic path enterprises should take to safe their AI:
Level merchandise can plug particular person gaps, however they’ll’t sustain with the velocity, complexity and interconnected nature of AI adoption. And extra importantly, they wrestle to ship the belief, consistency or scale AI transformation requires.
Many organizations are already experiencing AI adoption outpacing conventional safety instruments. Safety groups are below stress on three fronts:
- Danger – Shadow AI, unmanaged brokers and {custom} LLMs create new pathways for knowledge loss, mental property publicity and mannequin misuse.
- Price – Every new AI use case brings one more software, driving up license, integration and operations prices.
- Complexity – Fragmented controls throughout community, knowledge, id and software stacks create blind spots precisely the place AI is shifting quickest.
From a CIO or CISO’s perspective, this isn’t only a technical concern however the fault line beneath their whole AI agenda. CIOs are below stress to ship productiveness good points, value efficiencies and new AI-powered capabilities sooner than ever earlier than.
CISOs, however, see a parallel actuality: custom-built AI purposes that could be insecure by default, brokers that may act unpredictably, and a continuing threat that firm secrets and techniques or buyer knowledge may leak into third-party GenAI instruments.
If AI strikes ahead with out safety, the enterprise is uncovered. If AI slows down as a result of safety can’t sustain, the enterprise misses its transformation targets. That is why AI safety isn’t a characteristic; it’s the figuring out think about whether or not AI turns into a aggressive benefit or a strategic setback.
Gartner recommends the trail ahead as “an built-in modular AI safety platform (AISP) with a standard UI, knowledge mannequin, content material inspection engine and constant coverage enforcement.”
Gartner additional recommends prioritizing investments in two phases.
Section 1
Begin with AI utilization management to safe the consumption of third-party AI providers.
Section 2
Broaden into AI software safety to securely develop and run AI purposes.
Earlier than enterprises can safe how AI is developed, they need to first perceive how it’s already getting used throughout the group. The earliest dangers typically emerge not from the AI-enabled apps constructed in-house, however from the exterior generative AI instruments and copilots workers undertake, and infrequently with out the IT groups’ information.
That’s why we predict the report identifies AI utilization management as section one and why we suggest IT leaders begin with these quick inquiries to assess their group’s AI utilization.
- The place is AI really being utilized in my group?
- Which instruments, copilots and brokers are in play, and on what knowledge?
- How do I allow productiveness with out dropping management?
As soon as public generative AI use is known, the tougher problem emerges: Securing the AI apps and instruments that your group creates for itself. As fashions, brokers and pipelines transfer into manufacturing, the questions shift from visibility to integrity, security and scale.
Key questions that organizations should reply in section two embody:
- What AI purposes, fashions and brokers are my groups constructing, and the place do they stay?
- How do I handle the integrity, security and compliance of AI apps earlier than they attain manufacturing?
- How do I defend fashions and AI purposes from immediate injection, misuse or agentic threats?
- How do I scale AI innovation with out creating safety bottlenecks for builders?
Though organizations can separate the work round securing AI utilization and AI improvement, they don’t seem to be two separate issues. The identical group that wants visibility into workers utilizing public GenAI apps additionally wants to guard the AI purposes and brokers they’ve constructed as they transfer into manufacturing. A platform strategy is what permits shared insurance policies, shared guardrails and shared context throughout each side of the AI utilization and improvement equation.
That’s precisely the philosophy behind our Safe AI by Design strategy:
- Safe how GenAI is used with Prisma® Browser™ and Prisma SASE to find AI instruments in use, govern entry and forestall delicate knowledge from flowing into public fashions, all whereas maintaining customers productive with GenAI and enterprise copilots.
- Safe how AI is constructed with capabilities of Prisma AIRS™, equivalent to mannequin and agent safety, AI safety posture administration, runtime safety, automated testing with AI Purple Teaming, in addition to protection for agentic protocols, like MCP, securing {custom} AI purposes, brokers and pipelines.
Gartner identifies Palo Alto Networks as “the corporate to beat” of their newly launched report as of December 8, 2025: “AI Vendor Race: Palo Alto Networks Is the Firm to Beat in AI Safety Platforms.”
We imagine we’re the AI Safety Platform to beat as a result of:
- Palo Alto Networks product portfolio throughout community, edge, cloud and knowledge offers a robust basis for AI utilization visibility and management.
- The acquisition of Shield AI built-in industry-leading AI expertise and merchandise ensuing within the not too long ago introduced Prisma AIRS 2.0, which delivers complete end-to-end AI safety, seamlessly connecting deep AI agent and mannequin inspection in improvement with real-time agent protection at manufacturing runtime. The platform, constantly validated by autonomous AI purple teaming, secures all interactions between AI fashions, brokers, knowledge and customers. This offers enterprises the boldness to find, assess and defend their whole AI ecosystem, accelerating safe innovation.
- Complementing the platform, Unit 42®’s deep experience and Huntr’s bug bounty program, present safety thought management that instantly improves product effectiveness and menace intelligence. These applications assist us constantly uncover new assault patterns, misconfigurations and provide chain dangers distinctive to AI methods, in addition to feed these insights instantly again into the product roadmap.
- Our giant put in base and distribution channels create a flywheel for AI safety platform adoption and studying from our prospects and companions.
We additionally imagine that beneath the technical necessities is a deeper fact: CIOs and CISOs wish to transfer quick on AI, however they solely really feel secure doing so with a accomplice who has the dimensions, sign and endurance. That is the place our breadth, analysis depth and ecosystem matter.
Being early is a bonus, however staying forward requires humility and steady studying. Main means seeing what comes subsequent, and Gartner’s insights speed up our personal roadmap as we proceed to evolve.
- Simplifying the Expertise: We’re integrating capabilities throughout Prisma AIRS, Prisma SASE and Prisma Browser to make AI safety simpler to undertake, function and scale via Strata™ Cloud Supervisor as the only entry level.
- Going Deeper into the AI Engineering Pipeline: We acknowledge that securing AI should begin early within the growing atmosphere and ML pipeline, not simply at runtime. Our integrations with AI improvement instruments and code repositories will proceed to develop.
- Preserving Tempo with a Quick-Transferring Market: We’re investing in open requirements, partnerships and analysis, so our prospects don’t should chase each level resolution that seems. Palo Alto Networks can also be a contributing member to OWASP Requirements and Menace evaluation to assist create an {industry} normal on AI safety.
- Working Alongside Native AI Controls: Cloud suppliers and AI platforms are including their very own safety features. We intention to enhance, not substitute, these controls, offering unified visibility, superior safety and constant insurance policies throughout a fragmented AI panorama.
For us, being “the corporate to beat” just isn’t a end line. It’s a accountability to hear rigorously to prospects, adapt as AI evolves, and hold delivering sensible, built-in outcomes quite than remoted options.
In case you are a GM, CIO, CISO or AI chief attempting to make sense of a quickly crowding AI safety panorama, we imagine “GMs: Win the AI Safety Battle With an AI Safety Platform” is crucial studying.
Ultimately, the actual race isn’t about options; it’s about who helps enterprises speed up transformation safely, scale back threat and compete higher with AI they’ll belief.
Disclaimer: Gartner doesn’t endorse any firm, vendor, services or products depicted in its publications, and doesn’t advise expertise customers to pick out solely these distributors with the best scores or different designation. Gartner publications encompass the opinions of Gartner’s enterprise and expertise insights group and shouldn’t be construed as statements of truth. Gartner disclaims all warranties, expressed or implied, with respect to this publication, together with any warranties of merchantability or health for a selected function.
Gartner, AI Vendor Race: Palo Alto Networks is the Firm to Beat in AI Safety Platforms, By Mark Wah, Neil MacDonald, Marissa Schmidt, Dennis Xu, Evan Zeng, 8 December 2025.
Gartner, GMs: Win the AI Safety Battle With an AI Safety Platform, By Neil MacDonald, Tarun Rohilla, 6 October 2025.
GARTNER is a registered trademark and repair mark of Gartner, Inc. and/or its associates within the U.S. and internationally and is used herein with permission. All rights reserved.
Leave a Reply