From Scripted AI Agents to Agentic AI: What’s Changed in Customer Service?
30/6/26, 10:00 am
Customer service automation is entering a new phase.
For years, the chatbot had a fairly simple job: greet the customer, offer a few menu options, answer common questions and pass the conversation to a human agent when the issue became too difficult.
That model helped organisations reduce repetitive enquiries and gave customers another way to find information. It also exposed the limits of scripted automation. Many customers have experienced the familiar loop of selecting options, rephrasing questions and eventually repeating the same information to an agent.
The next phase of customer service is more ambitious. AI agents can understand intent, interpret context, draw from trusted knowledge, complete tasks across systems and work alongside human employees. This shift from scripted chatbots to Agentic AI is beginning to reshape how organisations think about service, experience and operational efficiency.
Scripted AI agents still have an important role to play. They are particularly valuable when journeys need to be predictable, repeatable and tightly controlled. The shift is not about replacing every scripted interaction with Agentic AI, but about matching the right form of automation to the right use case.
At a glance
- Customers increasingly expect service to be immediate, personalised and connected
- Traditional chatbots are useful for simple tasks, but often struggle with complex or changing customer needs
- Agentic AI introduces a more capable model, where AI can reason, act and escalate when needed
- Knowledge management is becoming critical to trusted AI-powered service
- Human agents remain essential for empathy, judgement and complex problem-solving
- The strongest outcomes come from orchestrating people, knowledge, data and workflows around better service experiences
Customer expectations have moved faster than service models
Customers think in outcomes, not channels.
A customer may start with a web form, move to a chatbot, call the contact centre and then follow up by email. From their perspective, it is one issue and one relationship with the organisation.
PwC Australia’s 2025 service centre trends commentary captures the pressure clearly: modern customers demand immediate, integrated and personalised services, while 80% consider experience to be as important as a company’s products.
Inside many service operations, that expectation is difficult to meet. Contact centres often rely on multiple platforms, disconnected knowledge sources and manual processes. Agents may need to search several systems or re-key information while customers wait.
Proof point: ServiceNow reported that Australians spent 123 million hours on hold in 2024, with almost eight in ten considering taking their business elsewhere after a poor service experience.
These conditions highlight why traditional chatbots, while useful, are no longer enough on their own.
The role and limits of scripted AI agents
Simple automation has value. Customer service is rarely simple.
Scripted AI agents are useful when the customer journey is predictable and the organisation needs a defined, deterministic path. They can support tasks such as answering standard questions, checking opening hours, confirming order status, guiding users through simple forms or managing repeatable service requests.
Their limits become clearer when the customer’s need is less linear. A person may describe the same problem in different ways, reveal new information partway through the interaction, need the organisation to check multiple systems or require judgement-based escalation.
The shift to Agentic AI is therefore not about replacing every scripted journey. It is about expanding what automation can safely handle when context, reasoning, orchestration and action are required.
Scripted AI agents and Agentic AI are not the same thing
| Scripted AI agents | Agentic AI |
|---|---|
| Follow defined flows and decision trees | Interpret context and determine the next best action |
| Work well for deterministic, repeatable journeys | Work well for dynamic, multi-step journeys |
| Provide tight control over approved paths | Operate within guardrails, governance and escalation rules |
| Are useful for standard enquiries and structured tasks | Can coordinate tasks across systems and workflows |
| Reduce effort in predictable service scenarios | Extend automation into more complex service scenarios |
| Remain valuable where risk, compliance or consistency require fixed outcomes | Require careful design where judgement, safety or accountability matter |
Agentic AI changes the role of automation
The shift is from scripted response to goal-driven service.
Agentic AI introduces a more capable model for service automation. Rather than responding only within a fixed script, an AI agent is designed to pursue a goal within defined boundaries. In customer service, that might mean helping a customer change an appointment, lodge a claim, update account details or troubleshoot a service issue.
Proof point: Gartner reported that 85% of customer service and support leaders planned to explore or pilot customer-facing conversational GenAI solutions in 2025.
Salesforce’s 2025 State of Service research found that AI has become the number two priority for service leaders globally, second only to improving customer experience. In Australia and New Zealand, Salesforce reported that AI is expected to handle half of customer service cases by 2027, up from 31% today.
With that pace comes a need for discipline. Organisations gain the most value when AI agents are designed around clear use cases, trusted information, escalation pathways and measurable service outcomes. Governance and human oversight are essential to using AI safely in customer-facing environments.
The next challenge is orchestration
AI is only as reliable as the knowledge behind it.
No AI agent can be more reliable than the knowledge it draws upon. If policies are outdated, content is duplicated or internal knowledge is scattered across systems, AI may simply expose the problem faster.
Modern customer service transformation depends on a strong knowledge foundation. AI agents, human agents and customers all benefit when self-service, assisted service and employee-facing knowledge draw from a consistent source of truth.
Human expertise remains especially valuable when the issue is complex, sensitive or emotionally charged. In those moments, AI can provide context and support while the human agent brings judgement and empathy.
Proof point: NiCE’s 2025 Global Happiness Index found that 72% of consumers report experiencing benefits from AI and automation in customer service. The finding suggests customers are open to AI when it improves the experience.
Knowledge and human judgement still matter
A standalone AI tool will only go so far.
To deliver meaningful outcomes, AI agents need access to customer context, interaction history, knowledge, workflows, business rules and escalation paths. They also need to hand over to human agents without losing the thread of the conversation.
Orchestration connects the moving parts of customer service: channels, data, knowledge, AI agents, human agents, workflows and escalation pathways. Without it, organisations risk creating another disconnected tool.
Questions organisations should ask now
- Where are customers experiencing the most friction today?
- Which interactions are repetitive, rules-based and suitable for automation?
- Which moments require human judgement, empathy or risk management?
- Is organisational knowledge accurate, accessible and governed?
- Can AI agents take action safely across systems?
- Are escalation pathways clear when automation reaches its limits?
- How will success be measured: containment, resolution, satisfaction, cost, employee experience or all of the above?
The future is intelligent, connected and human-centred
The chatbot era helped organisations understand the value of digital self-service. It also revealed the limits of automation that cannot act, reason or connect across the enterprise.
Agentic AI points to a more capable future. Customers can receive faster and more personalised support. Employees can be better equipped to resolve complex issues. Organisations can improve efficiency while maintaining the human judgement that service still requires.
Customer service has always been about solving problems. What has changed is the intelligence, speed and coordination now available to help solve them.
The chatbot answered questions. The AI agent helps get things done.
Move from AI interest to AI action
Explore how NEC’s Agentic AI capability helps organisations design intelligent AI agents that can understand intent, take action and support more connected customer experiences across channels and systems.
