Microsoft’s rollout of autonomous artificial intelligence (AI) agents marks one of Big Tech’s boldest pushes yet to unleash software that can independently handle customer service and business tasks with minimal human oversight.
The move, challenging Salesforce’s recent entry into intelligent automation, shows how enterprise software providers race to give businesses AI systems that can run portions of their operations, from scheduling meetings to resolving customer complaints to processing orders.
Industry experts say this shift could slash operational costs for early adopters while raising fresh questions about how companies should select and implement sophisticated AI tools that aim to replicate human capabilities.
“Autonomous AI agents will be game-changers for both customer service and operations,” JJ Lopez Murphy, head of data science and AI at Globant, told PYMNTS. “On the customer service side, they’ll act more like intelligent advisers, helping people get the help they need faster with better, more personalized responses. This means happier customers and less strain on support teams, all while cutting costs.”
AI agents are computer programs that can perceive their environment, make decisions, and take actions to achieve specific goals. They range from simple rule-based systems to complex models that can learn from experience and adapt their behavior, similar to how a chess program observes the board, evaluates possible moves, and chooses actions to try to win the game.
Agents of Change?
Microsoft’s autonomous AI agents, launching next month for Copilot, will let businesses automate routine tasks like researching sales leads, tracking supplier delays, and managing customer service inquiries. The system pulls data from Microsoft 365, Dataverse, and Fabric platforms to execute tasks independently.
Microsoft said that early pilot results from companies like McKinsey show time savings, cutting client onboarding processes from typical time frames to 10% of the expected duration. Microsoft is releasing 10 prebuilt agents targeting specific business functions in sales, service, finance and supply chain management. However, details on their reliability and limitations have yet to be public.
Agents like those offered by Microsoft could have specific customer service and commerce applications. The agents drive personalized recommendations, using AI to analyze user behavior and offer tailored suggestions, Mykhailo Maksymenko, head of salesforce practice at the digital consultancy firm Customertimes, told PYMNTS.
“This reduces wait times, enhances customer satisfaction, and allows human staff to focus on more complex tasks,” he said. “On the operational side, AI agents can automate daily processes like data management, supply chain optimization, and predictive analytics, driving cost savings and improving overall efficiency. As more companies adopt these agents, we may see a fundamental shift towards more seamless, 24/7 service models.”
In one recent example, Murphy pointed out that Target rolled out an AI tool to help employees quickly resolve on-the-job challenges, making day-to-day operations more efficient.
“AI agents like this can adapt on the fly — adjusting their approach, tools, and goals in real time, which is perfect for handling complex or unexpected tasks that customers throw at retail employees,” he said. “Imagine not having to try to hack the support chat so you get a human agent because it automatically solves your problems.”
Choosing an Agent
When selecting an AI agent system, Murphy cautioned that reliability varies significantly between platforms, with some being more prone to mistakes and faulty decision-making than others.
“So, picking platforms with solid reasoning skills and good self-checks is important,” he said. “It’s not like ‘one’ of them is hallucination-free, but the stronger the model, and the better you can control its surroundings with things like guardrails and other methods, the smoother the experience will be.”
According to Murphy, a critical factor is how well the AI platform connects with your current tools and software, ideally allowing for easy customization and piece-by-piece implementation.
“And, of course, you’ve got to think about costs — don’t go overboard on spending, but make sure you’re getting solid efficiency gains,” he said. “But the most important thing is educating your team on using AI agents properly and ensuring they understand data privacy and ethical considerations.”
