Editorial
In my last article, I discussed how AI companies are racing to release new applications and noted that the industry’s pace is unlikely to slow any time soon. Data from EY US found that “97% of senior business leaders whose organization is investing in AI report positive ROI from their AI investments.” Of those companies, 34% plan to invest $10 million or more next year — a directional increase from 30% six months ago in the first wave.
Another report from PwC noted that more than half of tech leaders say AI is fully integrated into business strategy, but only 30% say it is fully integrated into operations. The C‑suite is enthusiastic about AI, but the people developing, operationalizing and actually using these applications have a more nuanced view.
Software Professionals’ Views on AI
This time last year, my first blog for VKTR cited research from my own company’s generative AI data. Now, the 2025 State of Digital Quality in AI Report results are in, revealing that the disconnect between accelerating AI adoption and essential quality assurance practices persists. Based on responses from over 4,400 software professionals and consumers worldwide:
- 70% of organizations are actively developing AI applications, yet only one‑third employ red teaming to mitigate risks.
- 66% encountered a problem with Gen AI since January 1, 2025, including:
- Responses lacking detail (40%)
- Misunderstood prompts (38%)
- Biased outputs (35%)
- Hallucinations (32%)
- Clearly incorrect content (23%)
- Offensive content (17%)
Many organizations are investing in AI to streamline operations and reduce costs, and the good news is that more than half of software professionals report productivity boosts up to 74% from AI-powered coding tools. At the same time:
- 23% say their integrated development environment (IDE) lacks embedded tools.
- 16% are unsure if the tools are integrated.
- 5% have no IDE at all.
The tools of choice remain GitHub Copilot at 37% (down from 41% in 2024) and OpenAI Codex at 34% (up from 24% last year). Chatbots and customer support solutions top the list of AI‑powered applications being built (55%), while only 19% have begun building AI agents.
Related Article: How AI Is Reshaping Corporate Decision-Making — and What You Need to Know
The Consumer View
Outside the tech sector, AI usage looks very different. A recent Pew Research survey of more than 5,200 U.S. workers found:
- 55% rarely or never use AI tools like ChatGPT, Gemini or Copilot at work.
- 16% say at least some of their work is done with AI.
- 25% believe that while they’re not using AI much now, some of their work could be automated.
Convenience drives consumer AI adoption: a Cognizant survey reports that “frustration with the buying process, noted by 75% of respondents, will push consumers toward AI solutions that save them time.” At the same time, they are not ready to cede full control — most won’t allow AI to automatically reorder and pay for high‑value items without direct authorization.
Applause’s latest research found that consumers can be fickle:
- 30% swap one AI service for another.
- 34% prefer different Gen AI services for different tasks.
Demand for multimodal AI functionality has risen sharply: 78% say the ability to interpret text, images and other media is important, up from 62% last year.
Making Sense of the Numbers
It’s clear that AI investment continues apace, despite uneven integration and persistent quality concerns. As business leaders move AI from strategy to production, rigorous testing — especially expert‑led red teaming— will be key to operationalizing these systems safely and reliably. Consumers will expect AI applications to perform as well as existing tools and to protect their trust.
Organizations can win that trust by prioritizing digital quality and the customer experience. Embedding QA best practices and adversarial testing into every AI initiative will help bridge the gap between investment and actual usage — and deliver on the promise of enterprise AI.
Learn how you can join our contributor community.
