Artificial intelligence agents are redefining the potential of generative AI in business, elevating chatbots to new heights by handling complex tasks independently.

These autonomous AI agents can execute a wide range of activities, from confirming car rental reservations at airports to identifying and screening potential sales leads.

Major software providers, including Salesforce , ServiceNow , Microsoft , and Workday , introduced their own AI agents last year, touting their ability to streamline operations in areas like employee recruitment, sales outreach, marketing content creation, and IT management—allowing businesses to take a more hands-off approach.

If these AI agents deliver on their promises, they could provide the measurable return on investment many organisations seek from generative AI. Corporate technology leaders highlight their potential to reduce employee work hours or even minimise the need for additional hiring.

Phu Nguyen , head of digital workplace at Pure Storage , sees AI agents as a game-changer for empowering every employee across the data storage firm.

“Why should only executives have access to a ghostwriter for their emails or slide decks? Imagine if all employees had that capability,”

However, the rise of AI agents brings its own set of challenges, particularly in cybersecurity. According to Gartner, AI-driven decision-making is set to grow significantly. By 2028, at least 15% of daily business decisions are expected to be made autonomously by AI agents, up from virtually none in 2024. But with that growth comes risk—Gartner also predicts that 25% of enterprise breaches will be linked to AI agent misuse by the same year.

Below, we explore five companies that are already integrating AI agents into their products and operations, along with the lessons they’ve learned along the way.

Johnson & Johnson: Drug Discovery with AI Agents

At New Jersey-based Johnson & Johnson, AI agents are transforming the chemical synthesis process in drug discovery, streamlining what has traditionally been a labor-intensive task.

Chief Information Officer Jim Swanson explains that after identifying a promising pharmaceutical molecule, the next step is to assess its cost-effectiveness and reliability—a process influenced by countless variables, including temperature and reaction optimization.

The solution? An autonomous AI agent capable of determining the ideal timing for a solvent switch. This critical step involves replacing one solvent with another to crystallise a molecule, effectively bringing the drug to life. “We’re using agents now to analyse the variables and decide when to make that switch—and even execute it,” Swanson said.

Traditionally, J&J scientists would manually iterate through this process, ensuring optimal conditions were met. Now, AI agents, combined with traditional machine learning and digital twins (digital replicas of real-world processes), accelerate the workflow, saving both time and resources.

Despite the efficiency gains, Johnson & Johnson is cautious. Swanson emphasises the importance of managing risks associated with autonomous agents, such as the potential for biased outputs or hallucinations. While employees currently review the agents' outputs, the company is exploring ways to make oversight more systematic and robust. Given the highly sensitive nature of Johnson & Johnson's deployment of AI agents, the company may fall within the scope of the EU AI Act if it serves customers or clients within the European Union. Under the Act, their AI agents would likely be classified as high-risk systems, and depending on the specific capabilities of these agents, they could even fall under prohibited categories when the regulation takes effect on February 2, 2024.

However, J&J’s proactive governance measures play a critical role in mitigating these risks. By maintaining stringent oversight and human intervention, the company demonstrates a commitment to compliance and ethical AI use, effectively reducing the likelihood of breaching the EU’s regulatory framework. This careful approach ensures that innovation can thrive within a tightly regulated environment.

Moody’s: Financial Analysis Agents with Diverging Perspectives

At Moody’s, AI agents are becoming indispensable in financial research, enhancing the way the New York-based firm analyses data and generates insights.

Nick Reed, Chief Product Officer, explains that many research tasks, such as industry benchmarking and reviewing Securities and Exchange Commission filings, were previously outsourced to lower-cost regions. Now, some of this work is handled by autonomous AI agents, particularly through a collaborative system of multiple agents.

Moody’s has developed 35 specialised AI agents, each tailored for specific tasks like project management, and paired them with supervisory agents to create a sophisticated "multi-agent system." These agents are programmed with distinct instructions, personalities, and access to data, enabling them to process research and reach unique conclusions.

This setup is particularly valuable for analysing nuanced topics, such as the financial stability of a company that appears sound but faces risks like geopolitical instability. Agents can approach such issues from different perspectives, sometimes arriving at conflicting conclusions.

“It’s almost like how an individual processes information,” Reed noted. “What we’ve discovered is that agents excel when they focus on one task at a time rather than multitasking.”

This approach not only enhances the depth and reliability of financial analyses but also underscores the potential of AI agents to bring diverse viewpoints to complex challenges.

eBay: AI Agents for Coding and Selling

eBay is harnessing AI agents to streamline operations, from writing code to crafting marketing campaigns. The company also plans to introduce agents designed to assist buyers in finding products and sellers in listing goods.

To power these capabilities, the San Jose-based online marketplace developed its own “agent framework,” which leverages multiple large language models (LLMs) behind the scenes, explains Nitzan Mekel-Bobrov, eBay’s Chief AI Officer.

This framework acts as an orchestrator, determining which AI models are best suited for specific tasks—whether it’s translating code or suggesting code snippets. As the agents evolve, their autonomy will increase, enabling them to write code line by line in a way that mimics human developers, Mekel-Bobrov added.

The agents also learn over time.

“As employees interact more and more with the systems, it adapts to their specific preferences,”

he said, highlighting the potential for personalisation and efficiency gains.

eBay’s innovative approach demonstrates how AI agents can enhance both technical processes and customer interactions, paving the way for smarter, more adaptive systems.

Deutsche Telekom: Ask-Me-Anything AI Agent for Employees

Deutsche Telekom, the telecommunications giant with around 80,000 employees in Germany, has deployed an AI agent to assist its workforce with questions about internal policies, benefits, and product or service inquiries.

Known as askT, the agent is used by approximately 10,000 employees every week, according to Jonathan Abrahamson, Deutsche Telekom’s Chief Product and Digital Officer.

The company is also piloting askT’s ability to perform tasks on behalf of employees. For instance, an employee planning their next vacation can instruct askT to submit their leave request directly into the HR software system, streamlining administrative processes.

Cosentino: A ‘Digital Workforce’ for Customer Service

Spanish company Cosentino, renowned for its countertop surfaces and stone materials, has introduced a “digital workforce” of AI agents to address gaps in its customer service team, says Rafael Domene, the company’s CIO.

Cosentino treats its AI agents as digital employees, equipping them with foundational skills and onboarding them with training. These agents follow strict processes, and their actions are closely monitored to ensure compliance. “We know if they go off the rails,” Domene noted.

The AI workforce has completely taken over tasks previously handled by three to four customer service staff, such as clearing customer orders. This shift has allowed human employees to focus on more value-added service areas, demonstrating how AI can complement and enhance human resources.

Closing Remarks

The companies highlighted in this article, spanning industries from telecommunications to financial analysis, showcase how established organisations with robust infrastructure and significant capital can drive AI adoption at scale. With the resources to develop sophisticated frameworks, conduct rigorous testing, and implement critical safeguards, they are well-positioned to leverage AI agents while mitigating the associated risks. These capabilities allow them to innovate confidently, striking a balance between efficiency gains and responsible use.

However, for SMEs, the landscape is far more challenging. Limited budgets, fewer technical resources, and less experience with complex technologies often leave SMEs at a disadvantage when adopting AI. Without proper governance structures in place, the risks of deploying AI agents—ranging from biased outputs to cybersecurity vulnerabilities—can outweigh the potential benefits, making it harder for these organisations to compete.

At Digital Bricks, we specialise in bridging this gap. We understand that adopting AI agents isn’t just about implementing solutions—it’s about building a sturdy foundation that enables sustainable growth. Our approach emphasises both the technical and governance aspects of AI deployment, ensuring that businesses can innovate responsibly and effectively.

We guide SMEs through every step of the process, from identifying the right use cases for AI agents to designing governance frameworks that align with industry standards and regulations, such as the EU AI Act. By helping organisations implement clear policies for oversight, monitoring, and accountability, we empower them to minimise risks and maximise ROI.

Our comprehensive solutions address critical questions that every organisation must answer:

• How do we ensure our AI agents remain unbiased and transparent?
• What measures can we take to prevent misuse or ethical lapses?
• How do we govern AI systems in a way that aligns with future regulatory changes?

We believe that AI has the power to transform businesses of all sizes, not just those with deep pockets. By providing tailored solutions and governance expertise, we enable you to adopt AI with confidence, turning challenges into opportunities and laying the groundwork for scalable, sustainable innovation.

Contact us today at info@digitalbricks.ai to discuss your AI Vision.

‍