As artificial intelligence becomes more integrated into business infrastructure, the concept of “AI agents” is shifting from experimental technology to practical enterprise tool. The Salesforce World Tour held recently in Amsterdam focused heavily on the emergence of AI agents as a scalable solution for operational challenges faced across industries. While Salesforce’s Agent Force platform was central to the presentations, the broader message was about the changing nature of work and the growing role of autonomous systems in customer-facing and back-office functions.
AI agents, in this context, are not limited to chatbots or basic automation. They are systems designed to interpret goals, access structured and unstructured data, reason through decisions, and execute tasks — often in collaboration with human staff. These agents are increasingly being deployed across functions such as sales, customer service, logistics, and analytics, driven by improvements in large language models, integration with enterprise systems, and advances in data infrastructure.
Responding to Workforce Challenges
Several drivers are pushing companies toward AI-enabled agents. Workforce capacity constraints, rising customer expectations, and pressures on operational efficiency have all contributed to the shift. As noted during the keynote, frontline staff across sectors — including retail, healthcare, and customer support — are facing growing demand, often without proportional increases in resources.
AI agents offer one response: systems that can operate continuously, respond instantly, and take over repetitive or high-volume tasks. The aim is not full replacement of human roles but augmentation — freeing employees to focus on more complex or sensitive issues while AI handles routine inquiries or transactions.
Real-World Deployment: Fisher & Paykel and ReMarkable
Fisher & Paykel, a kitchen appliance company, provided one of the more detailed case studies presented at the event. The company implemented AI agents to streamline their customer support operations. In a demonstration, a customer interacting with a digital agent was able to request a refrigerator exchange, upload a photo of the installation space, receive product recommendations, and schedule a new delivery — all without engaging a human representative.
The agent also demonstrated the ability to reason through constraints, such as adjusting delivery times based on technician availability. Importantly, the system included oversight: tasks beyond the agent’s scope triggered a seamless handoff to human staff. Fisher & Paykel reported increased resolution rates and greater efficiency in their support operations as a result.
Another example came from ReMarkable, the Norwegian company behind the paper-like digital tablet. Facing rapid growth and demand for high-touch service, the company adopted AI agents to handle common customer interactions. According to the keynote, 35% of their customer base now interacts with an agent, allowing human staff to focus on more complex service needs.
System Requirements and Data Dependencies
Successful deployment of AI agents hinges on more than the AI model itself. As several speakers emphasized, the agents rely on access to accurate, relevant data. This includes product information, customer histories, logistics data, and other structured and unstructured content. Retrieval-augmented generation (RAG) — a method that allows the system to access external data sources in real time — was described as a key enabler.
Salesforce highlighted the importance of harmonized data infrastructure through tools like its Data Cloud, but the principle applies more broadly: the effectiveness of AI agents depends on data quality and accessibility.
Additionally, integration with business systems is essential. AI agents must not only generate responses but also take action — placing orders, updating records, initiating workflows. For this, APIs and process automation tools must be in place and properly configured. Reasoning engines within the agents are used to determine the sequence of actions required to fulfill a user’s intent.
Governance and Transparency
The keynote also acknowledged the risks and challenges associated with autonomous systems. Trust, transparency, and compliance were identified as ongoing concerns, particularly as agents are given access to sensitive data or decision-making authority.
To address this, Salesforce emphasized the need for monitoring tools and compliance frameworks. Their implementation includes visibility into an agent’s reasoning process — what actions were considered, what data was used, and what decisions were made. This traceability is intended to support internal audit functions and build confidence in agent behavior.
Beyond Customer Service: Enterprise-Wide Adoption
While many early applications of AI agents have focused on customer service, the technology is expanding into other areas. Use cases presented at the event included automated sales coaching, marketing campaign generation, quote creation, inventory checks, and field service coordination. One example showed a service technician using an AI-generated pre-work briefing and voice-to-text documentation to streamline repairs.
Another featured a merchandising strategist using agents to identify growth opportunities based on sales data and inventory forecasts. In each case, the AI agent was positioned as a collaborator — operating alongside human staff with the ability to process information and take action quickly.
A Gradual Path to Autonomy
Despite the enthusiasm, the transition to widespread AI agent use is expected to be gradual. The keynote outlined a three-stage approach: first, using agents to answer questions based on internal data; second, enabling them to take specific, defined actions; and third, allowing agents to proactively initiate tasks based on observed needs or opportunities.
Early adopters are encouraged to start with small-scale deployments, validate performance, and expand as organizational readiness and data maturity improve.
Looking Ahead
AI agents are unlikely to replace entire job functions in the near term, but their growing capabilities signal a significant shift in how work is distributed and performed within enterprises. Their ability to operate across platforms, handle routine tasks, and interact in natural language formats positions them as a useful addition to the digital workforce.
The rise of AI agents is not just a technological trend — it reflects broader changes in workforce strategy, data use, and customer expectations. As demonstrated by companies like ReMarkable and Fisher & Paykel, AI agents are already delivering measurable benefits, suggesting a continued move toward hybrid human-AI operations in the years ahead.
