News

The pace of change in AI has felt bewilderingly fast over the past 12 months, with new technologies emerging and seemingly being usurped on a weekly basis. For decision-makers, this can be a daunting challenge. However, the encouraging news is that AI development is largely iterative, each new tool builds on the foundations laid by its predecessors.This has brought us to the next phase of the AI revolution, Agentic AI. This latest development describes the development and implementation of autonomous software agents, grounded in Generative AI, that can make decisions and take action independently of human input. According to Gartner, by 2028, 33% of enterprise software applications will include AI agents, and 15% of work decisions will be made autonomously. Forward-thinking organizations are already using AI agents to uncover business value and achieve goals such as accelerating software development.Yet, just as Generative AI needs training data to be truly effective, AI agents need a clear understanding of business context. How can leaders ensure that AI agents comprehend how their businesses operate? The answer lies in Process Intelligence (PI). PI takes data from systems such as Enterprise Resource Planning (ERP) and Customer Relationship Management (CRM) software to track how events progress within an organization. It creates a dynamic, living digital twin of business operations, offering a holistic view of how work gets done. This makes it a foundational tool for implementing AI in ways that actually deliver value.Why AI agents?Agentic AI refers to autonomous ‘agents’ that can handle complex tasks independently. Many agents are armed with access to Large Language Models (LLMs), along with access to business-specific data (for instance, knowledge base articles or the order information). Employees can interact with many of them using natural language, asking them to then rapidly analyze business data to work out what the next step of a process should be, and even take follow on actions automatically.AI agents are not, however, a one-size-fits-all technology panacea that can solve every business problem right out of the box. For AI agents to succeed, they must be built to solve specific problems and they need insight into how the business really functions.This is where PI plays a critical role. It gathers together fragmented data from across dozens or hundreds of business processes, offering AI agents a ‘common language’ to understand events such as invoicing and shipping, and offering high-quality, timely data which can enable AI agents to make better decisions.With a ‘digital twin’ of business operations in hand, AI agents can analyze how processes truly impact each other across the whole business, and uncover opportunities to drive efficiency.Putting AI agents to workBusinesses are already creating AI agents built to harness the power of PI and seeing tangible results. One customer has worked with Celonis to develop an AI-driven inventory to track parts and materials. Within two months the AI tools had identified that many purchase orders were raised for spare parts that were already in stock as well as highlighting that a significant portion of spare parts were over eight years old.An additional AI Agent uses the inventory to optimize spare part availability for plant engineers, with users able to describe the parts they need using technical descriptions or common industry terms, eliminating the need for exact part numbers.In another case, PI and Agentic AI helped a company double the speed of software delivery by improving predictability and cutting stage waiting times by 30–40%. AI-driven tools pinpointed bottlenecks, offered predictive alerts, and suggested mitigations ranked by potential impact. Leaders could ask simple, natural-language questions to uncover delays and risks, using an AI copilot that translated complex data into clear, actionable insights.Why AI needs PIAgentic AI holds the potential to revolutionize enterprise operations, but its effectiveness depends on the quality of data agents have access to. PI ‘bridges the gap’ to provide AI with the input it needs, offering oversight of the totality of the business’s processes. PI is thus a vital tool for optimizing enterprise processes.Enterprise customers who try to improve their processes using AI without the vital insights from PI all too often fail. In fact 89% of business leaders globally we surveyed recently said that giving AI the context of how their business runs is crucial if it is going to deliver meaningful results.That is why we believe there can be no effective enterprise AI without PI. Process intelligence is integrated into live systems, so even when systems change, it offers AI agents real time access to the current state of processes. Think of it like the mapping data for a GPS.Without a map, you’re just following a line on a blank screen. You won’t know why you were turning left and it would be all too easy to take a wrong turn. Similarly, Process Intelligence gives AI agents the essential context to navigate business complexity reliably.A smarter futureAgentic AI is set to become increasingly central to enterprise success. But its impact depends on access to timely, accurate, and contextual data. Process Intelligence provides this foundation—enabling AI agents to drive meaningful change across business functions, from software development to finance.The message is clear: Agentic AI needs the right data, and the right context. That’s exactly what Process Intelligence delivers.We've featured the best AI chatbot for business.This article was produced as part of TechRadarPro's Expert Insights channel where we feature the best and brightest minds in the technology industry today. The views expressed here are those of the author and are not necessarily those of TechRadarPro or Future plc. If you are interested in contributing find out more here: https://www.techradar.com/news/submit-your-story-to-techradar-pro

Xreal detailed Project Aura at AWE 2025It will have a massive 70-degree field of viewIt will also be tethered to a spatial computing puck running Android XRWhile we'd suspected that Android XR would be a key component of Google I/O 2025, we couldn’t have predicted some of the partners Google announced that it would be working with, which include the gadget makers at the top of our best smart glasses list: Xreal.As promised by Xreal at I/O, it has taken to the Augmented World Expo 2025 stage in Long Beach California to provide us with new details on its Project Aura glasses, and it’s shaping up to be one impressive device.For me, the most important detail is that the device will apparently boast a 70-degree field of view, which is absolutely huge.The 50-degree field of view of the Xreal One already felt large, and the 57-degree Xreal One Pro is a noticeable step up size-wise (you’ll need to wait a little longer for our full review). 70-degrees will be massive.The field-of-view upgrade suggests – Xreal hasn’t confirmed this yet – that the Aura specs will borrow the Xreal One Pro’s new optic engine (and perhaps even upgrade it further) including its flat-prism lenses, as one of its key advantages is that it enables a greater FOV .This optic system comes with other upgrades as well, which could help to make the Android XR glasses much easier to use all day as you walk around.(Image credit: Google)Another interesting tidbit is that these specs – like Xreal’s other glasses – are tethered, meaning they’re powered by an external device which they’re connected to via a cable.We already knew Aura wouldn’t be standalone, but Xreal has revealed that the new compute device shipping with Aura won’t just be your standard phone, or the Xreal Beam Pro.It’s something all-new, running Android XR, and powered by a Snapdragon chip from Qualcomm – which seems to be making all of the Android XR processors.Xreal isn’t abandoning its own chipset however. Aura itself will sport an upgraded X1S chip that’s a “modified version of X1 with even more power under the hood.” The X1 chipset is what’s inside the Xreal One and Xreal One Pro specs.A new X1 chipset is coming (Image credit: Future)Xreal has yet to confirm if it will sell the puck and glasses separately, but if it does then I'll be interested to see what that decision means for its approach to the upgradability of its tech going forward.At the moment you can pick up a pair of Xreal glasses and a Beam spatial computer as a bundle, and then upgrade either or both over time. Newer glasses offer better visuals and audio if that’s your main concern, while the new Beam Pro offers improved processing and spatial features.This is a less wasteful and generally more affordable design philosophy, as you only need to replace the one component that’s holding you back. However, as I mentioned, Xreal has yet to confirm if it will sell the puck and glasses separately. Its current wording calls Project Aura “one solution made up of the wired XR glasses and a dedicated compute device” suggesting they might also be one complete, non-upgradable package.As for a launch date, Xreal is still keeping us mostly in the dark, though it has said Project Aura is coming in 2026, so we hopefully won’t be waiting for too long.Xreal One Pro dead on arrival?The Xreal One Pro (Image credit: Future)Following this announcement some fans are starting to wonder if their Xreal One Pro purchase was a good idea – if they'd waited a year or so longer and they could have snagged an Xreal Android XR setup instead.I’ll concede that for some Xreal One Pro purchasers waiting may indeed have been the better approach, but I think others can rest easier, as while the Aura and One Pro will likely share similarities I suspect they’ll be very different devices.For a start, while Xreal’s glasses are often at their best with the Beam Pro add-on, it isn’t required. You can use the specs with a range of USB-C compatible devices, and even many HDMI devices with the right cables.Based on Xreal’s descriptions so far Project Aura isn’t just a wearable display for entertainment; it’s a complete spatial computing package with all the nifty Android XR features we’ve been shown.This won’t just mean that Aura’s purpose is different from Xreal’s other glasses; I expect its price will be very different too.Right now an Xreal Beam Pro and Xreal One Pro would cost you $848 / £768 (before factoring in any bundle or limited-time discounts). For what sounds like it will be greatly improved hardware I imagine Project Aura will cost closer to $1,000 / £1,000, if not more.The Xreal Beam Pro (Image credit: Xreal)And remember, you can buy the Xreal One Pro separately for just $649 / £579.Better tech is always on the horizon at any given time, but this (for now) doesn’t look set to be a repeat of the Meta Quest Pro / Meta Quest 3 fiasco, which saw the latter, far superior product launch at less than half the price of the former.Instead Project Aura looks set to be more of a diagonal shift, with new hardware boasting better specs and a different purpose.If you want to wait for Project Aura you absolutely should, as you might also be tempted by many of the other Android XR, Meta smart glasses, and new Snap spectacles set to be launching in the next year or so. But choosing not to wait isn’t a bad option either, and the Xreal one pro certainly isn’t dead on arrival like some might think.You might also likeMeta Quest 4: 5 things I want to see from Meta's next VR headsetI took my Meta Quest 3 on a 3,000-mile flight so you don’t have toForget Android XR, I've got my eyes on Vivo's new Meta Quest 3 competitor as it could be the most important VR headset of 2025

You’re no doubt familiar with shadow IT — the practice of employees using software, applications and other tech tools that aren’t sanctioned by IT. And if IT doesn’t know about something, they can’t regulate it or defend against it. Clearly, this creates a massive security risk and headaches for both IT and security.Now, with generative AI tools flooding the workplace, that headache is turning into a full-blown migraine.The rush to adopt AI productivity tools has opened a Pandora's box of security vulnerabilities that most organizations are completely unprepared for. These tools are expanding existing visibility gaps while simultaneously creating a constant stream of new ones.Invisibility — in plain sight“Patchy” (pun very much intended) visibility is only part of the problem. There’s a widespread awareness of the threat posed by unregulated tools, but there’s a startling gap in translating that awareness into concrete readiness. According to Ivanti’s 2025 State of Cybersecurity Report, 52% of IT and security professionals view API-related and software vulnerabilities as high- or critical-level threats. So, why do only 31% and 33%, respectively, consider themselves very prepared to address these risks? It’s the difference between theory and practice.Making that shift to readiness is easier said than done, given the widespread and elusive nature of shadow IT practices. Software that employees use, including shadow IT, ranks as the number one area where IT and security leaders report insufficient data to make informed security decisions — a problem affecting 45% of organizations.Let that sink in: nearly half of security teams are operating without visibility into the applications running within their own networks. Not good. At all.The Gen AI multiplier effectGenerative AI has created a perfect storm for the proliferation of shadow IT. Employees eager to boost productivity are installing AI tools with little thought to security implications, while security teams struggle to keep pace.The ubiquity and ease of access to these tools mean they can appear in your environment faster than traditional software ever could. A text summarization tool here, a code generation platform there — each creating new pathways for data leakage and potential breaches.What makes this particularly dangerous is how these tools operate. Unlike traditional shadow IT applications, which often store data locally, generative AI solutions typically send corporate data to external cloud environments for processing. Once your sensitive information leaves your controlled environment, all bets are off.The root of the shadow IT problem…This isn’t a manifesto on regulating employee behavior. It’s genuinely understandable, at least to me, why employees would seek out tools that help them do their work more efficiently. That is to say, shadow IT isn’t always done out of malice, but rather because something is lacking in the organizational structure.Specifically, data silos between security and IT teams create perfect conditions for shadow IT to flourish.These divisions manifest in a few different ways, for example:Security data and IT data are walled off from each other in 55% of organizations62% report that siloed data slows security response times53% claim these silos actively weaken their organization's security postureWhen IT lacks visibility into security threats and security lacks visibility into IT operations, shadow IT thrives in the gaps between.…and how to solve itAddressing the shadow IT challenge, particularly in this AI-centric era, requires a totally different approach from what IT and security teams might have tried in the past. Instead of attempting to eliminate shadow IT entirely — a likely futile effort — organizations need to build frameworks that provide visibility and control.Breaking down those data silos that separate IT and security teams is a critical first step. This means implementing unified platforms that provide comprehensive visibility across the entire attack surface, including shadow IT and the vulnerabilities it creates.With proper integration between security and IT data, organizations can move from reactive firefighting to proactive defense. They can identify unsanctioned AI tools as they appear, assess their risk levels and implement appropriate controls — all without hampering the productivity gains these tools offer.Of course, dismantling silos is an oversimplified directive. There needs to be an ongoing culture shift where employees no longer feel the need to engage in shadow IT practices covertly. Employers should listen to employees about what tech-related barriers they face. Employee-preferred tools should be evaluated for potential inclusion. Employees must be trained on risks and understand how their choices directly impact business outcomes.Micromanagement is certainly not the solution, nor is AI itself the problem. AI is a reality of our current workplace, and a lot of good stems from many of the new AI tools. The problem comes when employers fail to dismantle silos, tackle visibility gaps, bring shadow IT into the open and proactively prepare for the attack vectors that come with these tools.Ignoring the problem will not make it go away. As generative AI continues to gain prevalence and capability, the problem will only worsen.We've featured the best online cybersecurity course.This article was produced as part of TechRadarPro's Expert Insights channel where we feature the best and brightest minds in the technology industry today. The views expressed here are those of the author and are not necessarily those of TechRadarPro or Future plc. If you are interested in contributing find out more here: https://www.techradar.com/news/submit-your-story-to-techradar-pro