What Tasks Will Be Automated in Future Businesses?

In today’s technologically advancing world, artificial intelligence (AI) disrupts almost every sector, reimagining how business is conducted. The ubiquitous influence of AI is not only confined to digital marketing or legal analysis, but its tendrils are reaching into virtually every aspect of commerce. From automating accounting processes to streamlining sales outreach to creating dynamic presentations, AI integration is revolutionizing the work landscape.

“I honestly can’t think of many areas of business that won’t be affected by AI at some point in the future,” says Dr. Craig Froehle, a professor of operations, business analytics, and information systems at the University of Cincinnati Lindner College of Business.

“That said, I think a lot of jobs will have relatively subtle effects in the near future, whereas other jobs will be more overtly and broadly affected. Take agriculture, for example. AI is helping forecast weather conditions and markets to assist farmers with the task of deciding what to grow and when to plant. But the physical processes of preparing land, planting seeds, tending to crops, and harvesting have been less affected, at least in the short run. In contrast, artists and translators are two jobs where professionals are already today feeling the pressure that generative AI models have started to exert on their professions over the past year or two.”

The field of medicine is another place where AI and automation have had a significant impact: “The Mayo Clinic Platform integrates a large-scale repository of patient and care process data with a suite of AI tools and algorithms to help healthcare organizations provide better care. While Mayo Clinic is probably further down that road than most healthcare systems, Cincinnati Children’s Hospital Medical Center has the Artificial Intelligence Imaging Research Center, which develops new AI-based algorithms and tools to help radiologists and physicians make better use of the complex imaging data generated by advanced modalities,” shares Dr. Froehle.

Keep reading to learn more about what jobs and tasks will be automated, the benefits of automation, the challenges of these technological changes, and ethical considerations.

Meet the Expert: Craig Froehle, PhD

Dr. Craig Froehle is a professor of operations, business analytics, and information systems at the University of Cincinnati Lindner College of Business. He also holds a faculty appointment in the UC Department of Emergency Medicine.

After working in the engineering industry and founding a dot-com startup, he moved to academia to pursue his research and teaching passion: operations management, with a special focus on healthcare delivery and technology-enabled services.

2026 Update: Craig Froehle on the State of AI in Business

In 2026, BSchools.org sat down with Dr. Craig Froehle again to discuss how the AI landscape has shifted since our original conversation in late 2023. A lot has changed, and some things, he noted, have barely moved at all.

“I reread the article, and it is amazing how far the technology has come, and yet how little progress has been made on how we cope with, manage, legislate, and regulate these things,” he said.

Adoption Has Been Faster Than Expected, But Not in the Way Many Predicted

One of the clearest shifts since 2023 has been the broad adoption of AI across organizations of every size and industry. What surprised Dr. Froehle is not just the pace of adoption, but where it has and has not happened.

“I think the adoption curve and the integration curve in terms of integrating AI elements into back office work processes and platforms have actually been faster than I expected,” he said. At the same time, the trend toward companies building their own proprietary models has been weaker than anticipated.

The reason, he explained, is that major AI providers have made their platforms so capable and accessible through APIs and other tools that building in-house often does not make financial or performance sense: “For a lot of companies, the big providers have also realized that they benefit from offering versions that are siloed so that it’s still running the best in class model, but the data aren’t escaping the organization.”

The exception is organizations with strict privacy or intellectual property requirements, such as hospitals and law firms, which are investing in tools that keep sensitive data internal. Even so, Dr. Froehle believes the integration challenges those industries are working through now will largely be solved within a few years.

Agentic AI and Multimodality Are Now the Norm

Two developments that have reshaped the AI landscape since 2023 are the rise of agentic AI and the normalization of multimodality. Where AI models once handled text primarily, it is now standard for them to process images, audio, and video interchangeably. “You just kind of assume that everything is everything. Information is information, no matter what form it’s in,” Dr. Froehle said.

Agentic AI, in which models can autonomously carry out multi-step tasks rather than simply respond to prompts, is quickly moving into enterprise software. Dr. Froehle cited a Gartner projection that by the end of 2026, roughly 40 percent of major enterprise applications will include task-specific AI agents, up from around 5 percent just a year prior.

Reasoning Models Have Raised the Bar on Reliability

Another major development is the rise of reasoning models, which go beyond predicting the next word in a sequence and instead work through problems in multiple internal steps before producing a response. The practical effect, Dr. Froehle said, is a meaningful reduction in hallucinations and a much higher ceiling on what AI can reliably do.

He offered a simple illustration. A few years ago, asking an AI model to count the words in its own answer would reliably produce a wrong number. Today, the best models not only get it right, but they also identify the most efficient possible answer. “It actually figured out what the most parsimonious answer could be,” he said. He also described a recent moment where a model, when gently pushed back on a factual claim, independently went through a multi-step verification process and returned with a corrected answer and an explanation of how it confirmed the correction. “That’s a huge improvement,” he said.

Hallucinations still occur, he noted, and the degree varies significantly by platform. But the overall trajectory is toward greater reliability, driven largely by these reasoning architectures.

Customer Service AI: Capabilities Are Ahead of Public Trust

Customer service was one of the areas flagged in the original article as ripe for automation, and Dr. Froehle says the technology has largely delivered. The sticking point now is trust. “We’ve so long relied on a person on the other end of the line to be helping us that the trust that a chatbot is capable is going to lag the capabilities,” he said.

Part of the problem is structural. Unlike in the European Union, where the EU AI Act requires public-facing AI to disclose that it is not human, US companies are under no such obligation. That transparency gap, Dr. Froehle suggested, makes it harder for consumers to calibrate their expectations or build confidence in automated systems. “Chatbots don’t have intention. They don’t have ethical obligations. There’s no desire to help you the way a customer service agent should arguably have. Ultimately, the only source of trust we’re going to be able to have in these automated tools is their efficacy and service quality. And if that remains poor, then that trust is going to be hard to develop.”

The US Regulation Gap

On the question of regulation, Dr. Froehle was candid. While the EU has moved forward with risk-based frameworks and disclosure requirements, he does not see significant US regulation on the horizon in the near term.

“Something truly shocking would have to happen,” he said, pointing to a hypothetical high-profile failure, such as an AI-related aviation incident, as the kind of event that might push legislation to the forefront. He noted that cases of individual harm, even serious ones, have not been sufficient to generate that level of public demand.

He does, however, see one scenario that could change the political calculus: visible, systemic unemployment. “I think the big impetus will come when we start seeing directly attributable unemployment,” he said. “If you start seeing systemic increases in unemployment and you can easily point the finger at AI, then you could get lots of angry people yelling at legislators to do something.”

Entry-Level Jobs Are Already Eroding

That employment concern is not hypothetical. Dr. Froehle pointed to entry-level programming and data analytics roles as sectors where the pipeline has already thinned dramatically. “Entry-level programming jobs have essentially dried up,” he said. The downstream consequence is a problem the industry has not yet solved: if people cannot gain foundational experience in junior roles, it is unclear how the next generation of mid- and senior-level professionals will develop.

“There’s a lot of local optimization that’s creating a nationwide sub-optimization,” he said. “Each individual company might be doing what’s best for it, but across all industries, we might be in a very bad way in four or five years if there isn’t anybody behind the people who retire or leave.” He added, with some irony, that if the answer to that gap is simply more AI, the question becomes whether the field enters a self-reinforcing spiral.

Physical AI and Robotics: The Next Frontier

Perhaps the most significant development Dr. Froehle flagged that was not on the radar in 2023 is the rapid acceleration of physical AI, meaning robots capable of operating in real-world environments. Advances in AI architecture have given robotics a new trajectory after years of incremental progress.

He pointed to Hyundai Kia’s manufacturing operations as a case study, with plans to deploy thousands of robots by 2030, and noted the growing presence of Chinese robotics companies as evidence that this is becoming a global race. “As of right now, we have robots working in factories. They are not widely deployed, but by 2030, you’re probably going to see some shift where a role that used to be a person has parts of it taken over by a robot,” he said.

The implications for business automation are significant. Jobs that seemed safely human because they required physical dexterity and navigation of unpredictable environments—think HVAC technicians, repair workers, and manufacturing roles—may not remain insulated for as long as previously assumed.

AI Literacy: Awareness Is High, Knowledge Is Uneven

One area where Dr. Froehle expressed measured concern is AI literacy. Adoption is increasing on two tracks: people actively seeking out AI tools, and people using AI without realizing it, through features like advanced autocorrect that have quietly shifted from rule-based to AI-powered systems. “Your autocorrect is AI-powered now, and you didn’t realize there was a shift,” he said.

What worries him, particularly in his role as an educator, is that widespread awareness has not translated into widespread understanding of how to use AI ethically, reliably, and effectively. He sees a gap in K-12 education, specifically, where large-scale AI literacy efforts have not yet taken hold. “Awareness of it is high. Knowledge of it, though, is still highly variable.”

Looking Back: Our 2023 Conversation on AI and Business Automation

Jobs and Tasks Most Likely To Be Automated

Several job roles and tasks show a high potential for automation as AI capabilities continue to evolve: “Jobs that rely heavily on routine have traditionally been the easiest to automate. Whereas a lot of automation in the past has been through procedural scripts and rule-based systems, new generations of AI models will help cover a wider range of tasks and functions,” says Dr. Froehle.

“Jobs that require some integration of various pieces of information in a fairly systematic way, such as drafting up a standardized legal document or generating an operational dashboard, are relatively straightforward today and likely to be automated.”

While there are many jobs and tasks that are already automated, in the future, there will be many more, as well as a marked improvement in the ones that already exist: “Moving ahead, much better language models, such as those used for translation, transcription, and customer support, will increasingly replace some parts of some workers’ jobs. On the software engineering side, automatically documenting, and sometimes even fixing, code has already shown to be feasible at previously unattainable levels of quality and scale,” explains Dr. Froehle.

He continues, “Healthcare, customer service, and a lot of information industries have already embraced AI and started using it in impressive ways. Many hospitals have AI running in various capacities, either home-grown or off-the-shelf. Radiology, for example, has long been a hotbed of applications in computer vision. Many online customer service and call center functions employ chatbots to help customers with routine questions, saving their human capacity for more complex or ambiguous situations. One of my doctoral students is researching how such chatbots might best be integrated into customer service systems to maximally help customers while taking as much strain as possible off the human agents.”

Some jobs will be much harder to automate. “On the contrary side, those tasks that are going to be least automatable involve significant amounts of critical thinking and innovation relative to highly technical or complex bodies of knowledge, training, or experience,” says Dr. Froehle. “It is truly hard to say what lies in store for us ten years from now.”

Impact on Workforce

The impacts of automation and AI on the workforce can be viewed through multiple lenses. On one hand, the implementation of AI can lead to job displacement. As noted earlier, roles involving routine tasks that can be automated, such as data entry clerks, manufacturing, and warehouse jobs, are at risk. A 2023 study from the McKinsey Global Institute found that by 2030, up to 30 percent of the hours currently worked in the United States could be automated.

On the other hand, AI and automation also have the potential to create new job categories and amplify human capabilities. For instance, AI can take over mundane tasks, allowing employees to focus on higher-level, more strategic responsibilities. This can lead to increased job satisfaction, productivity, and innovation.

“Like any large-scale technology-induced shift, jobs will change. They changed when steam power enabled at-scale manufacturing. They changed when electrification allowed factories to run 24/7. They changed when computerization eliminated repetitive, mundane information tasks and calculations,” shares Dr. Froehle.

“What is harder to predict is how jobs will change because AI tools are evolving so rapidly right now. Last spring, I prepared a new curriculum for a course that was to happen in August and September. By July, some of what I had prepared was already obsolete. That rate of change makes forecasting much into the future extremely difficult, if not impossible. Today’s workers should be prepared to upgrade their skill sets faster and more frequently than any generation of workers before them. And short of AI replacing the vast majority of workers entirely, something I truly do not expect, I don’t see that trend slowing down.”

Ethical Considerations of Business Automation

One of the most significant issues with AI and business automation is the ethical problem of algorithmic bias, where AI systems may inadvertently reinforce existing prejudices due to flawed data or programming. This can lead to unfair outcomes in hiring, lending, or law enforcement: “How will we ensure that unjust biases that exist in historical data don’t infect future models? If they do, then the decisions made by those models will be no better or fairer than the decisions we’ve made in the past,” warns Dr. Froehle.

“They’ll just be faster and cheaper. While that’s a big advantage, not straining out those biases whenever possible would be a huge missed opportunity to use technology for real and necessary good.”

Aside from baked-in biases, there are other ethical considerations with AI-created work. “Another issue is the intellectual property rights of those who generated the data we’re using to train these large AI models. While some can be trained exclusively on operational data, that is to say, data that is generated simply by executing business processes, a lot of training requires the use of either data scraped off the internet, which is actually owned by someone, or data labeled by workers who may not be paid a living wage. It’s similar to raw materials in a supply chain from questionable or problematic sources,” explains Dr. Froehle.

Benefits of Business Automation

The automation of business operations through AI and other methods presents a multitude of benefits: “The primary benefits are first faster/cheaper outputs, and second, insights that human-based analysis likely would not generate alone,” shares Dr. Froehle.

“Automating a decision-making process, such as discriminating between conforming and faulty parts on a production line, makes them cheaper than when they’re done one-by-one by people. Labor is just really expensive, largely because it’s typically not very scalable and also partly because humans make lots of mistakes, get bored, get sick, etc.”

He continues, “Another benefit is less about saving money and more about creating a competitive advantage. Suppose your AI-based protein-folding model can generate new pharmaceutical formulas better than your competitors’ products. In that case, that’s really powerful, and doubly so if it can do it faster than their scientists can.”

Another benefit of automation is the ability to scale operations. As businesses grow, the tasks that need to be performed also multiply. Automation allows companies to handle this increased volume without a proportional rise in costs or staffing.

Automation can also lead to more consistency in business operations. Automated tasks are performed the same way each time, ensuring uniformity and reducing the probability of errors. It can also provide valuable data for businesses, which they can collect, organize, and analyze, leading to better decision-making.

Challenges of AI and Business Automation

While the advantages of AI and automation in business are apparent, implementing these technologies is not without its hurdles. “There are lots of challenges. One is the rapid rate of change. Businesses are already finding it hard to hire and retain AI talent. Keeping a stable regiment of AI scientists and engineers available to tweak existing models and develop new ones is an ongoing challenge,” says Dr. Froehle.

Implementing AI and automation requires a significant investment, not just financially but also in terms of time and resources for staff training. This is especially true for small and medium-sized enterprises that may lack the necessary resources: “Companies have to figure out how to maximize their return on investment. These systems can be quite costly to implement, and determining where in the company’s many processes and functions these models can justify themselves quickly can be hard. I expect companies that focus on how a specific system might add value will generally outperform those businesses that are constantly chasing the latest tech,” says Dr. Froehle. “Both these challenges are made worse by that rapid rate of change we’re seeing.”

Another challenge that businesses will face with automation and AI is data privacy. As AI systems become more prevalent in commerce, they collect vast amounts of personal data, raising questions about how this information is stored, used, and protected. This has led to the emergence of regulations such as the General Data Protection Regulation (GDPR) in Europe, adding a new layer of complexity for businesses utilizing AI.