Crystal Ball 2026: Prediction Roundup Part 4

We conclude the Crystal Ball Series close to where it began with this piece, with even more observations from our friend and occasional programming collaborator Frank Balonis of Kiteworks. But there are more SMEs here, including Operant AI CEO Vrajesh Bhavsar, who has more security insights where AI adoption and systems integration are concerned, and AI and human resources insights by Stacey Richey, who has an thought-provoking title at Smartcat: global VP of people.

Indeed, the supplementing of the human workforce in manufacturing with artificial intelligence, sometimes even “digital workers” for a litany of repetitive tasks, will be in our journalism at Smart Industry much in the next 12 months. The Crystal Ball Series was just the beginning.

See also: AI that augments the workforce … and doesn’t replace it

Frank Balonis, chief information security officer and senior VP of operations, Kiteworks

Data sovereignty becomes a standard clause in supply chain contracts. Manufacturers will start embedding data residency, routing, and logging obligations directly into OEM–supplier contracts instead of treating them as after-the-fact security add-ons.

With 46% of respondents citing GDPR-focused readiness and 34% citing cross-border transfer mechanisms as priority frameworks, the pressure to prove where data lives and how it moves is already mainstream.

Form data overtakes file shares as the top IP risk. As more quality, warranty, engineering change, and service workflows move into web and data forms, manufacturing will see sensitive IP increasingly exposed through structured submissions rather than shared files.

While 64% of organizations already secure web and data forms today and another 26% plan to by 2026, only 25% rank forms as a top data exchange risk—showing that many still underestimate how much sensitive information flows through them. In manufacturing, where portals and forms are central to supplier and customer interactions, that blind spot will give attackers a head start.

Vrajesh Bhavsar, Operant AI CEO

The autonomous agent heist. By mid-2026, attackers will exploit AI financial assistants to drain millions of bank accounts simultaneously without any user interaction. As banks and consumers increasingly rely on AI assistants to move money, pay bills, and manage finances 24/7, attackers will exploit a critical blind spot: AI agents inherently trust the data they ingest.

By hiding malicious instructions in everyday content like receipts, invoices, PDFs, and financial news feeds, attackers can silently manipulate these agents into authorizing fraudulent transfers. These attacks require zero clicks, bypass all phishing defenses, and appear completely legitimate because the AI uses its own approved credentials.

The trust exploitation crisis. Late 2026 will see trusted third-party AI tools turn malicious after routine updates, creating a supply chain attack that compromises thousands of enterprises simultaneously. Health care organizations will discover their AI assistants have been silently copying patient records for months through what appeared to be normal database queries.

Richard Copeland, CEO of Leaseweb USA

Trusted execution environment technology will reshape distributed compute and multi-cloud architecture. In 2026, trusted execution environment (TEE) technologies will finally move from “interesting concept” to real-world game-changer.

See also: Why AI is quickly becoming essential manufacturing infrastructure

We’re going to see organizations secure memory and hardware in a way that simply wasn’t practical before, which opens the door for decentralized compute in a very big way. Companies will be able to safely split compute across multiple clouds, regional providers, and even on-prem environments, instead of keeping all their workloads under one hyperscaler’s roof.

What is interesting to note here is that the shift isn’t driven by budgets or hype, but by behavior. When you can secure workloads at the hardware level, you’re suddenly free to architect systems around business needs instead of who owns the data center.

AI becomes truly agentic—replacing tasks, not people—and drives a new phase of cloud repatriation. AI is no longer just a tool for optimization. In 2026, agentic AI starts replacing full workflows, and that shift will separate companies that understand how to use AI from those that fight it.

The real impact isn’t that AI replaces jobs but that it replaces the tasks people shouldn’t be doing in the first place—the repetitive, time-sucking operations that drain teams. Organizations that lean into agentic AI will run faster, make decisions earlier, and redirect people into work that moves the business.

As AI becomes more embedded in day-to-day operations, more companies will realize that complexity and cost are pushing them away from the hyperscalers. They’re seeing outages, noisy-neighbor issues, unpredictable billing, and environments so complex that one failure cascades through the whole stack.

AI workloads, especially GPU-heavy ones, run better, and more cost-effectively, when the infrastructure is simpler, more transparent, and built for their exact workloads. That’s why 2026 will be a major year for cloud repatriation back to regional providers and bare-metal platforms built for performance.

Stacey Richey, global VP of people, Smartcat

ROI of AI: In 2026, a working understanding of how to use AI effectively will become a basic requirement for many roles. The organizations that see the most value from AI will be the ones that strengthen the culture and systems that guide how employees use these tools in their day-to-day work.

See also: Why AI is quickly becoming essential manufacturing infrastructure

Readiness gap and leadership responsibility: Many companies are already seeing a gap between the speed of AI deployment and employees’ ability to use it effectively, and I expect that gap to widen next year unless leaders put the right support in place. The companies that do this well will be the ones that focus as intentionally on their people and workflows as they do on the technology itself.

Distinguishing substance from hype in 2026: Much of the hype in HR tech focuses on AI replacing the function or transforming talent management overnight. The real substance in 2026 will be far more practical.

HR leaders will need to balance using AI to streamline their own routine work with their broader responsibility to enable the organization. AI will help teams streamline routine work, strengthen judgment through better insights, and support more consistent decision-making across the organization.

In 2026, AI’s role in the business will only continue to grow, and a working understanding of how to use these tools effectively will become a basic requirement for many roles. Employees will be expected to apply AI appropriately, recognizing where it supports their work and where its limitations require human judgment.

Many organizations are already seeing a gap between the speed of AI deployment and employees’ ability to use it effectively. I expect that gap to widen next year unless companies invest in training and set clear expectations and guardrails for internal AI use.

The results companies achieve from AI will depend not just on the technology, but on the culture and systems that guide how work gets done. Leaders across the organization will be responsible for helping their teams build these capabilities, with HR providing the standards, guidance, and resources that support that growth.