A new survey of enterprise use of artificial intelligence released Thursday by the Boston Consulting Group shows the technology has stalled in its deployment, and the most-hyped new area, “agents,” has seen low adoption within companies so far.
The obstacles to greater AI usage are primarily a lack of training, limited access to tools, and insufficient encouragement from management, the group states. These obstacles are raising concerns about job loss and lack of oversight, the Boston Consulting Group found.
Also: 15 new jobs AI is creating – including ‘Synthetic reality producer’
“AI is now part of our daily work life — but frontline employees have hit an adoption ceiling,” according to the report, “Momentum Builds, But Gaps Remain,” based on a survey of 10,635 employees at companies around the world from May 13 through June 4.
The study included employees at companies with more than $5 billion in revenue, as well as at those with less than $500 million, so it was not limited to large enterprises.
The survey covered three classes of employees: so-called frontline workers, managers, and “leaders” in organizations.
Use of AI is “strong” overall, the survey finds, with 72% of workers using the technology regularly, defined as “several times a week or daily.”
AI usage has hit a wall among frontline workers
While 51% of frontline employees use AI — up from 20% in 2018 — that figure is actually down one point from 52% last year.
The study authors identified three impediments to greater usage. The first is training: just over a third of employees report receiving adequate AI training, meaning “trained on the skills needed for AI transformation.”
The group advises that employees with five or more hours of training were more likely to use AI regularly, as were those who received in-person training (as opposed to virtual) and those who had access to an AI “coach” — a human who can advise them.
The factors “significantly boost employees’ confidence in AI and improve the quality of AI-enabled work outputs,” the study advises.
Also: Tech prophet Mary Meeker just dropped a massive report on AI trends – here’s your TL;DR
On the second point, almost 40% of employees “say their company is not supplying the right tools,” the survey revealed, though it did not specify which tools. Over half of those employees said they resorted to bypassing corporate restrictions to use other AI tools “when corporate solutions fall short,” and as many as 62% of Gen Z employees said they did so. That poses security risks for corporations, the group advises.
The third impediment was lack of support from the top, the group found. People were more likely to use AI regularly “with clear leadership support” for the technology. The survey found that “only 25% of frontline employees experience it,” referring to C-suite support.
AI usage is not advanced or sophisticated
Aside from individual use, the survey reveals that AI usage is rather basic, not advanced or sophisticated.
Of the companies surveyed, 72% said they were deploying ChatGPT, Microsoft Copilot, and French startup Mistral’s generative AI.
Also: AI could erase half of entry-level white collar jobs in 5 years, CEO warns
That percentage is well above the 50% of firms that reported they are “reimagining” corporate functions by redesigning “processes.” Only 22% of companies said they were developing something entirely new with AI.
Merely deploying ChatGPT is not ambitious enough
The group warns that merely deploying ChatGPT and similar rudimentary solutions is not ambitious enough.
“According to BCG studies, companies that create the most value with AI focus 80% of their investment on Reshape and Invent, in a few core processes.”
The firm states that companies which train workers and receive support from senior management are more likely to use AI to reimagine or invent processes rather than simply deploy a chatbot.
Also: Don’t be fooled into thinking AI is coming for your job – here’s the truth
“As a result, employees at those companies,” the authors say, “save more time” and “shift to strategic tasks,” and are more likely to “think their company will make better decisions thanks to data.”
Reimagining processes, however, has a downside. When asked if their job will be replaced by AI, 41% said yes, a number that rose to 46% in companies focusing on reimagining processes.
The sense of job loss
That prompted the survey authors to conclude, “Employees in companies reshaping their workflows feel most vulnerable to job loss — reinforcing the need for clear communication and proper upskilling.”
The sense of job loss was even higher in regions where AI is more prevalent. In parts of the “global south” — such as India, Spain and the Middle East — where usage is highest, respondents answered affirmatively by 48%, 61% and 63%, respectively, when asked if “their job will certainly or probably disappear entirely in the next ten years.”
Also: Open-source skills can save your career when AI comes knocking
Those surveyed reported other concerns about AI generally within their organizations. Concerns included “decisions taken without human oversight” (46%), “unclear accountability when mistakes occur” (35%), and “bias or unfair treatment introduced” (32%).
Also somewhat stalled are the much-ballyhooed agentic AI technologies: only 13% of companies currently have agentic AI in their “workflows,” despite 77% of respondents believing they “will be important in the next three to five years.”
Employees don’t understand AI agents
Over a third of respondents do not fully understand agent technology, the group found. “Employees see potential, but don’t fully understand AI agents,” they write.
The countries most active in putting agents “into their workflows,” according to their employees, are Brazil, India, Spain, the US, and the UK.
Also: Your next job? Managing a fleet of AI agents
“Three out of four employees believe AI agents will be critical to future success — but only 13% say they’ve seen them integrated into workflows,” said David Martin, the global lead of people & organization at Boston Consulting Group, in an email to me.
“That tells us we’re at the very beginning of the adoption curve. What we’ve found is that when people understand how agents work, they stop seeing them as threats and start seeing them as teammates. Education, not just experimentation, will be the unlock.”
The BCG report echoes other studies of the market that have emphasized obstacles and pitfalls in corporate AI deployment. For example, a January report from consulting firm Deloitte found that the majority of companies are not ready to put generative artificial intelligence into production.
Want more stories about AI? Sign up for Innovation, our weekly newsletter.
Leave a Reply