Metrigy research found most companies are confident in network infrastructure for AI, but concerns about security vulnerabilities and latency remain among IT leaders.

Metrigy’s research over the last few years has consistently revealed a willingness among significant percentages of companies to implement AI features and tools as they seek ways to bolster employee efficiency, reduce existing or avoid new costs, drive revenue and otherwise improve the business. But, we wondered, how confident are these companies that their network infrastructure was up to the job of all this AI demand?

We put this question to the IT and CX leaders who participated in our AI Technology Foundation & Strategy: 2026-27 global research study of 759 companies, conducted in March and April. Most, it turns out, don’t seem to have overwhelming concerns about their company’s network infrastructure for handling their AI and automation use cases.

Continue reading at nojitter.com.

Metrigy’s Matt Craig, senior director of product development, shares the Q2 view of our AI Sentiment Index. Tune in to discover how the index and sub-indexes have changed from their Q1 baselines, and what the top takeaways are from his analysis.

What the Q2 2026 Metrigy AI Sentiment Index Tells Us

Americans rate AI more favorably for what it does today than for where it is heading, and our second-quarter 2026 data puts a number on that split: a current-sentiment score of 102.2 against a future-expectations score of 100.6. People like what AI does for them this week. They are less sure they like where it is taking them.

We have been tracking this with the Metrigy AI Sentiment Index, a quarterly read on how 1,000 U.S. adults feel about AI across their personal lives, the economy, and society at large. We ask how they feel about AI right now and where they think it is headed over the next year or two, then roll those answers into an index. The first quarter of 2026 set the baseline at 100, so every number is a comparison against that starting point.

In the second quarter, the overall index nudged up to 101.4. That is a small move, and one quarter does not make a trend, so I am not going to pretend a 1.4-point bump means the world has changed. What is interesting is not the size but the split underneath it, that 1.6-point gap between how people feel about AI today and where they think it is going. They feel noticeably better about the present than the future.

The present is winning

That gap is the whole story this quarter, and once you start looking for it, it shows up everywhere. The share of people who say AI has made them personally better off rose from 44.8% to 47.2%. Regular users (daily or weekly) ticked up from 32.8% to 34.5%. Worry that AI is moving too fast actually eased a bit.

The reason is pretty human when you think about it. The good stuff people get from AI is concrete and happening right now: a thing they figured out in an afternoon, a task that was stuck and suddenly is not. The scary stuff lives in the future and stays abstract: job losses at scale, deepfakes nobody can catch, systems making big calls with no human in the loop. It is easy to worry about what AI might do to your industry in five years and still be glad it saved you two hours yesterday. Both feelings are real, and people are clearly holding both at the same time.

The long-range view backs this up. When we ask where AI lands in ten years, people split almost evenly: about a third expect widespread problems, a third expect growth and opportunity, and a third expect a messy mix. That barely budged from the first quarter, even after a few months of huge AI investment, new model launches, and nonstop headlines. A ten-year outlook that refuses to move is anchored to something deeper than the latest product demo.

AI moved to the front of the task

This one from our detailed report deserves more attention than it gets. Back in 2024, the common framing was AI as proofreader, the thing you run your finished work through at the end. That is over. Today, 29.7% of users say they bring AI in at the very start of a task, to figure out where to even begin. Another big chunk uses it for first drafts. Only 4.5% save it for polishing at the end. AI went from being the spell-check at the finish line to the thing you open before you have written a single word.

That shift matters for anyone building these tools. A lot of AI interfaces are still designed around the old one-shot model: type a question, get an answer, done. But most engaged users have moved past that. They treat AI like a back-and-forth partner, refining over several exchanges. The product that assumes a single prompt is increasingly out of step with how people work. And the higher earners, the ones already most optimistic, pull AI in earliest of all, which only widens the productivity gap between them and everyone else.

People can tell when AI is just a button

The other tidbit worth pulling out: people are getting tired of AI sprinkled on top of products they already use. Only 29.6% of folks call the new AI features showing up in their apps genuinely useful. More than 40% see them as either marketing hype or an outright annoyance.

That is a real signal. For two years, the move was to bolt an AI feature onto everything and put it in the press release. Users have caught on. They can tell the difference between AI that does something for them and AI that just adds another button to ignore. Apple took heat through most of 2025 for moving slowly on AI, and by early 2026 some of that caution was starting to look less like falling behind and more like reading the room. Feature-creep AI may start costing goodwill instead of earning it.

Who is actually optimistic

The positive feelings are not evenly spread. Income is the biggest dividing line by a wide margin, with high earners far more positive than low earners on basically everything. Education tracks closely behind. Gender and age matter too, just less: men and people aged 25 to 44 lean more positive than women and the 55-plus crowd. Sketch the composite optimist and you get a higher-income, highly educated knowledge worker in a tech-friendly city. Which makes sense. That is exactly the person whose daily work slots neatly into what AI is currently good at.

What I am watching next quarter

The one number I will be staring at is that present-versus-future gap. It could close if the doom headlines never quite arrive and people relax about the future. It could close from the other direction if the day-to-day magic wears off as novelty fades. Or it could stretch wider, with real usefulness pulling current sentiment up while the future worries dig in. The index is built to tell those apart, and that is the point: the longer we track it, the more each quarter tells us.

This is the short version. For the fuller picture, check out the summary report, which digs into the topic deeper. Current Metrigy clients can go deeper still with the detailed report and the underlying data analyzing the workplace side of all this, where unsanctioned “shadow AI” use is fading fast as companies finally catch up, plus the specific tools people reach for, what frustrates them, and where the data points for the AI business. But the headline is simple enough: people are using AI more, liking it more in the moment, and still not sold on the ending.

Second-quarter 2026 results show current sentiment climbing while 10-year expectations hold flat, exposing a present-positive, future-anxious divide.

June 17, 2026 — Metrigy today released the second-quarter 2026 results of its AI Sentiment Index, finding that U.S. adults are growing more comfortable using artificial intelligence (AI) in daily life even as their long-term expectations for the technology remain uncertain.

The Overall AI Sentiment Index rose to 101.4 from the first-quarter 2026 baseline of 100, lifted by a Current AI Sentiment Index of 102.2. The Future Expectations AI Index had little impact on the overall index, staying nearly flat at 100.6. The 1.6-point gap between current and future sentiment is the most significant feature of the quarter, with current personal experience pushing present sentiment up and persistent concerns about jobs, privacy and misinformation holding future expectations lower.

The broader survey data tracks in the same direction. The share of respondents saying AI has made them personally better off climbed to 47.2% from 44.8%, with the share of those considering themselves to be regular AI users growing to 34.5% from 32.8%. Misinformation, data privacy, and job displacement each remained primary concerns for more than half of respondents, and the 10-year outlook barely moved, splitting roughly evenly among continuous growth, widespread problems, and mixed results.

“People may genuinely worry about what AI will do in five years and still report that AI made their week easier. The data shows direct experience driving optimism while uncertainty keeps future expectations in check, and that divide is exactly what businesses need to design and market against,” said Matt Craig, senior director of product development at Metrigy.

The findings also point to a hardening divide by income and education. Among respondents earning more than $125,000, 68.5% feel personally better off from AI, compared with 19.9% of those earning under $20,000—a 49-point gap. 

Concern is highest among the most informed. Counter to the assumption that skepticism stems from unfamiliarity, 72.1% of master’s degree holders cite data privacy as a concern, compared with 33.3% of those with only some high school education. The most-exposed segments are the most aware of what could go wrong, undercutting the idea that the public simply doesn’t yet understand AI.

Consumer patience for surface-level AI is also wearing thin. Only 29.6% of respondents call new AI-powered features in existing products genuinely useful, while 41.4% view them as marketing hype or an active hindrance, a signal that future launches will be judged on what the AI actually does rather than on its presence.

The Metrigy AI Sentiment Index reflects the views of 1,000 demographically representative U.S. adults, sampled to be balanced across age, gender, household income, region, and educational attainment, with first-quarter 2026 establishing the baseline at 100. The full second-quarter 2026 detailed report is available from Metrigy at metrigy.com.

About Metrigy: Metrigy is an innovative research and advisory firm focusing on the rapidly changing areas of artificial intelligence, customer experience, and workplace collaboration—along with several related technologies, including knowledge management, security, and workforce engagement management. Metrigy conducts research with business leaders, consumers, and technology vendors. Our analysts deliver strategic guidance and informative content, backed by primary research metrics, indexes, and analysis, for technology providers and enterprise organizations.

Media Contact:
Elly Hertz, Sr. Director of Content Marketing and Design

Phone: 877-638-7449

Cell: 815-735-7059

[email protected]

Enterprise collaboration hardware and platforms are adjusting to the newest tools for conferencing and digital work.

This week’s InfoComm conference in Las Vegas will draw tens of thousands of attendees from across the A/V spectrum. While the show covers a broad range of technologies, enterprise collaboration is poised to play a major role, with featured keynotes from both Cisco and Microsoft.

Heading into the event, here are the major trends and topics that I am watching:

Continued buildout of the Microsoft ecosystem

Continue reading at nojitter.com.

As companies integrate artificial intelligence (AI) deeper into their daily operations, a new risk emerges. Many organizations rely heavily on external vendors and third-party frontier models to power their customer experience and internal workflows. This deep integration creates a critical question for IT and business leaders. What happens if a vendor changes its pricing, alters its data privacy policies, or simply fails to meet performance expectations?

Model dependency is one of the topics we explore in our new AI Technology Foundation & Strategy: 2026-27 research study of 759 companies globally. In the study, we found that far more companies are worried about the technical difficulty of switching models than the legal risk of being stuck with a model. Untangling a massive language model from an enterprise data architecture requires significant effort, and without a strategy, a necessary transition could easily become a costly operational nightmare.

To combat this risk, forward-thinking organizations build formal backup routes. Today, 53.4% of companies maintain a formal exit plan to guide them if they decide to switch the way they use AI today. Furthermore, implementing these strategies heavily correlates with overall business success. Nearly 65% of this study’s success group, determined by measured improvements from AI’s use, maintains a formal exit plan—showing that the most successful organizations prepare for change long before a crisis occurs.

An effective AI exit plan requires more than just a legal escape clause in a vendor contract. It demands a technical roadmap that ensures business continuity. When successful companies build these fallback strategies, they typically rely on several key tactics.

Elements included in an exit plan are:

Ultimately, building an exit plan requires cross-functional collaboration. IT teams must design the flexible architecture, while legal and compliance teams must ensure contracts guarantee the safe return of proprietary data.

You invest heavily in AI to drive revenue, boost efficiency, and improve customer satisfaction. Protecting that investment means preparing for every scenario. By developing a comprehensive exit plan today, you can ensure that your organization remains agile, secure, and ready to adapt to whatever changes the technology market brings tomorrow.ut as a strategic engine for job creation, workflow redesign, and unprecedented employee engagement.

Automated customer sentiment evaluations can help organizations understand their customers better, replace manual tasks and boost revenue by uncovering buying signals.

Every day, businesses leave critical customer insights buried inside support tickets, live chats and phone transcripts. This sinkhole exists because organizations have long relied on post-interaction surveys to capture customer sentiment. These surveys have a key flaw: They lack adequate responses from all customers except from the most delighted or deeply frustrated outliers. Despite this deficiency, the post-interaction survey remains the go-to mechanism for gathering customer sentiment for most companies, according to Metrigy research.

However, forward-thinking contact center operations are tapping into the power of AI inferred sentiment analysis to get themselves out of the mire. Today, inferred sentiment analysis uses large language models and context-aware processing to discern emotional tone of customer interactions. Essentially, inferred sentiment delivers a customer sentiment score for every touchpoint without requiring the customer to fill out a single survey.

Since Metrigy began tracking inferred sentiment in early 2023, we’ve seen a big bump in use, from 15% to 45% in early 2025. But the real story is in the reported improvements in core operational metrics from its use.

The benefits of inferred sentiment analysis

Continue reading at techtarget.com.

New research from Metrigy shows what jobs are now in demand — and which ones are disappearing — thanks to automation and productivity improvements.

You don’t have to look far to find people who are concerned AI is going to take their jobs. And depending on what their jobs are, their fears could be well-grounded.

But broad-scale concerns about AI taking jobs is overstated, according to Metrigy’s AI Organizational Best Practices 2026-27 global study of 759 companies. In fact, when asked how AI is affecting jobs at each research participant’s company, 43.7% say AI has created jobs, 23.2% say it’s eliminated jobs and 42% say AI has not affected jobs either way.

Among those research participants that selected the option “both created and eliminated jobs,” (because both can be true), 66.3% say AI has created more jobs than it has eliminated. Tracking job changes is vital for high school, college, and early-career employees — or to anyone in a career that is showing high negative impact of AI. Education, retraining, and certifications are vital to success for both employees and employers. Our research is showing some expected changes, but others that are bucking early predictions about AI’s impact on the job market.

Contrasting views from employers and employees

Continue reading at nojitter.com.

Thanks to their task-specific focus and local processing, this AI model makes sense for a variety of CX use cases.

When considering AI for customer experience (CX) use cases, matching the right model to the right task is a given—and, increasingly, this means figuring out if a small language model (SLM) is the better choice over a large language model (LLM).

The difference isn’t so much in how an SLM works compared to an LLM, but rather in their scale, efficiency, and where they’re deployed. In a nutshell, LLMs are massive, which means they require intensive compute power only available in the largest data centers. While they excel in answering complex, multi-step customer inquiries or generating highly nuanced, long-form content, they carry high latency and significant computational costs.

Continue reading at nojitter.com.

Every vendor wants to leverage their platforms as the system of record, but success may depend on interoperability.

While AI has been the buzzword of the last several years, “system of record” has emerged as the new core of AI product messaging as software vendors seek to differentiate themselves from one another, and ensure the value of their AI capabilities.

What is a system of record?

Continue reading at nojitter.com.