A 2025 study of online retailers found that implementing artificial intelligence into business operations improved conversions by up to 16.3 percent. Another recent study credited AI with increasing productivity by 2.4 percent across over 500 Japanese enterprises. A third study found that AI business solutions were associated with better decision-making.
Yet that third study also pointed out that, for many businesses, implementation remains a problem. As its authors write, “The most frequent barriers include employee resistance, high costs, and regulatory ambiguity. Respondents indicate that organizational factors are more significant than technological limitations.”
In other words, the problem isn’t the model. It’s the inexpert onboarding of AI resources into business processes.
According to Akhil Verghese, founder and CEO of Krazimo, a groundbreaking AI company for enterprises, businesses should follow a few best practices to ensure they get the most out of AI-powered solutions.
Businesses struggle to incorporate AI business solutions
Verghese has seen many companies achieve high ROI after implementing AI-powered automations. “In particular, AI-driven solutions have shown that they can do a great job with lead management, customer service, and sales assistance,” he says.
Yet he has also seen many businesses struggle for multiple reasons. “First and foremost, some business leaders are deceived about what they’re really getting,” he says. “They purchase a product that has been marketed as an intelligent AI solution that will transform business processes. But they realize later that they’re just chatbots.”
According to Forbes, “To be truly agentic, applications must be capable of completing complex tasks and long-term goal-oriented planning, with minimal human intervention.”
Business research advisory company Gartner refers to this deceptive marketing as “agent washing” and recently estimated that, out of the thousands of products on the market that purport to have AI capabilities, only about 130 actually do.
According to Verghese, many business leaders also rush the process of incorporating AI into their operations. “Going too fast introduces real vulnerabilities,” he warns. “Security and privacy concerns aside, if you don’t take the time to adequately define expectations, success criteria, and clear metrics for evaluation, you’ll never be certain of the efficacy of your deployed solutions.”
In addition, employees are often reluctant to integrate AI platforms into their work.
Employee resistance to AI tools
As Verghese explains, business executives tend to be more excited about AI-driven automation than the rest of their team members. “Business leaders see the demos and the promise,” he says. “They spend more of their time looking to the future, evaluating trends, and strategizing.”
Meanwhile, the people who have to do the actual work frequently discover that the AI generates new challenges and even slows them down.
“Employees who are tasked to use generalized tools for custom tasks are finding that, while AI helps a lot with writing tasks and general problems, it’s not solving a significant percentage of their daily work without focused effort in building customized solutions,” Verghese says.
For Verghese, there’s also a psychological factor to employees’ resistance. “Executives are the ones making the decisions to implement AI; they have agency,” he explains. “Employees are the ones having the change done to them. In general, I’ve found that people are more optimistic about changes they control than changes forced upon them.”
Supporting employees adequately as they deploy AI models is therefore Verghese’s first best practice for businesses.
How to support employees while onboarding AI platforms
To help employees start automating their tasks, Verghese advises purchasing the right AI solutions. “Buying a general enterprise subscription just isn’t going to work,” he says. “Companies need real customization.”
For this reason, he advises slowing down and taking a mindful approach. “Just because you’ve come across a neat AI system doesn’t mean it’s the right one for your company,” Verghese says. “Optimal business outcomes require intelligent solutions that have been carefully strategized from the beginning. Otherwise, you won’t boost productivity; you’ll just stress out your employees.”
Another reason companies need to slow down before launching an AI initiative is to ensure their data is in the right form. “Companies that understand AI understand the value of quality data,” Verghese explains. “They’ve cleaned, classified, and labeled information to make it usable. This process is real work that takes time and should be approached carefully. The quality of your implementation depends on it.”
For Verghese, companies also aren’t doing enough to train and equip their employees to use AI. “You can’t plug an AI in and expect it to immediately start taking over routine tasks, much less improving efficiency and streamlining operations,” he explains. “Deployments take time. Phased deployments with strict oversight and clear accountability are key to adequately testing solutions.”
Tips for using AI automation
According to Verghese, the implementation process is handled best in distinct stages. In the first, the AI attempts to do the same thing as human staff members. A qualified professional then compares the results to track the algorithm’s progress.
Once the machine demonstrates competence in the vast majority of cases and refrains from causing any harm, the process can advance to the next stage. The AI is allowed to contribute more with constant oversight from a human staff member, who provides feedback about its performance.
“The transfer to full autonomy needs to wait until the AI has earned it,” Verghese says. “Even then, there are times when it would not be appropriate to use AI tools, such as when legal ramifications would come into play. Or it could make costly mistakes, such as issuing a reimbursement to a customer.”
That said, Verghese notes that AI becomes most productive when it is ubiquitous within an organization. “Companies doing things right incorporate AI everywhere — sales, customer service, marketing, data analytics, coding, and more,” he says.
Finally, Verghese recommends approaching AI with a growth mindset. “The companies that get the most out of their AI investments understand the value of initial failures,” he says. “They cultivate a culture of experimentation. An AI project failing should never be seen as a career risk.”
Supercharge business intelligence, accelerate and optimize workflow with scalable AI
Improve customer service. Provide actionable insights. Predict future trends. AI can do all this and more — but only if it is implemented properly.
“When you innovate, you want to empower your team, not set it back,” Verghese says. “New solutions help build better businesses. Take the time to launch them the right way.”
