AI Appreciation Day exposes split over workplace risks

Artificial intelligence executives are using AI Appreciation Day to spotlight contrasting views on the technology's rapid spread across workplaces. Leaders from Kore.ai, Catalogic Software and Cynomi outlined both the promise of AI and the operational risks of current deployment.

At Kore.ai, Chief Strategy Officer Cathal McCarthy described a shift in how businesses view human contribution as AI tools become part of everyday work. He said the technology's progress over the past two years has redefined what organisations value in employees.

"Artificial intelligence has come a long way in the past two years, and it's made us rethink what makes humans valuable in this AI world. We have evolved from task executors to orchestrators, freed from mundane work to concentrate on the skills that matter most: judgment, empathy, creativity, and curiosity. AI is helping us become more humane again by giving us more time to think critically, ask questions, make better decisions, and work more effectively with each other. What's particularly interesting is that AI is the first technology that learns with us. As people become more skilled at using AI, the tech becomes more effective as well, creating a continuous cycle of improvement. Rather than replacing human expertise, AI amplifies it. AI Appreciation Day is a great time to appreciate this extraordinary partnership between people and tech," said McCarthy, Chief Strategy Officer, Kore.ai.

McCarthy's comments reflect a view of AI as a complement to human judgment rather than a substitute. His focus on orchestration and collaboration aligns with broader corporate experiments with generative AI in customer-facing and internal roles.

Others placed greater emphasis on risk, governance and clear objectives. Catalogic Software Chief Executive Officer Ken Barth warned that many businesses still approach AI adoption without an operating plan or defined outcomes.

"Most companies that report disappointing results from AI did not pick bad tools. They skipped the step where someone decides what the tool is for. Licences spread across teams, and six months later finance asks what the spend produced. Nobody can answer, because nobody defined the question. The companies getting durable value work backward from a specific task. A support team measures how long agents spend summarizing tickets, deploys a model to draft those summaries, and tracks whether resolution time drops. The scope is narrow, the baseline is measured, and the result is checkable. This approach also surfaces AI's real boundary quickly: models compress information well (summaries, extractions, first drafts) and handle judgment calls poorly when the context sits outside their view. Responsible adoption comes down to operating decisions someone has to own. Who reviews output before it reaches a customer? What data leaves the building in prompts? When the model is wrong, how does anyone find out? The answers are boring: review steps, approved-tool lists, retention terms, spot checks. Boring is the point. A company with those answers can expand AI use with confidence; a company without them accumulates risk at the speed of adoption. Restraint does not mean waiting. It means deploying first where errors are cheap and verifiable, then expanding as verification improves. The tools are the easy part. The decisions were always the work," said Barth, Chief Executive Officer, Catalogic Software.

Barth's remarks highlight a growing split between companies that embed AI in tightly scoped workflows and those that roll out tools across teams without clear measurement. His argument points to an emerging discipline around AI operations, echoing earlier waves of software adoption in which finance and compliance teams demanded traceable results.

In high-stakes environments, cybersecurity specialists are putting those ideas to the test. Cynomi Chief Executive Officer David Primor said AI is already reshaping how managed security providers structure their services and margins.

"One of the biggest advantages AI brings to cybersecurity is its ability to help service providers scale their expertise. At Cynomi, we've built an agentic AI platform that acts as a force multiplier for MSPs, MSSPs, and vCISO firms by automating time-intensive tasks such as risk assessments, remediation planning, compliance, and reporting. This enables our partners to deliver more strategic, proactive, and consistent cybersecurity services while expanding their businesses. We've seen partners achieve up to four times greater capacity, up to 30% higher margins, and up to 60% revenue growth, demonstrating how AI can strengthen both cybersecurity outcomes and business performance. AI isn't replacing cybersecurity professionals, it's helping them extend their expertise where it's needed most," said Primor, Chief Executive Officer, Cynomi.