Automation has become an integral part of customer service. Chatbots, voicebots, IVR systems, self-service tools and artificial intelligence applications help companies handle growing volumes of requests, reduce response times and improve operational efficiency.
However, automation does not automatically translate into a better customer experience. The real competitive advantage comes from the ability to use interaction data to understand which activities to automate, which to optimise, and when to hand the customer over to a human agent.
This is the central message of CallMiner’s 2026 CX Landscape Report, based on a sample of 700 senior contact centre and customer experience decision-makers in the United States, the United Kingdom and Ireland, France, Germany and South Africa.
Automation Is Widespread, but Not Always Intelligent
The report highlights that 99% of organisations now use some form of automation in customer interactions. While 63% consider themselves highly automated, only 23% say they combine a high level of automation with continuous optimisation based on customer experience insights..
This distinction is fundamental.
A company may have chatbots, IVR systems and automated workflows, yet still offer complex journeys, irrelevant responses or repetitive transfers between channels and agents. In these cases, the technology is in place, but it is not truly guided by customer intelligence.
The risk is creating a customer experience that appears efficient from an internal perspective but is frustrating for customers. Typical warning signs include:
- Customers having to repeat the same information multiple times;
- Systems failing to recognise the context of a request;
- Chatbots being unable to understand complex situations;
- Transfers to human agents happening too late;
- Automation being applied to interactions that require empathy, judgement or problem-solving skills.
For this reason, automation should be assessed not only in terms of cost reduction or the number of contacts handled without human intervention, but above all in relation to the quality of the outcome for the customer.
From Data to Action
Organisations are collecting increasing amounts of information: conversation transcripts, operational data, post-interaction surveys, real-time feedback, online reviews, social media comments, and chatbot and voicebot performance indicators.
According to CallMiner, the problem is therefore not a lack of data. All the organisations surveyed collect customer experience data, and 66% use automated processes to analyse it. Yet 68% say they are not making the most of their CX data.
The gap becomes particularly evident when moving from analysis to tangible improvements:
- 42% use insights to improve satisfaction, loyalty and retention;
- 38% use them to identify and resolve issues before they escalate;
- Only 30% use them to improve the performance of automated interactions;
- Just 22% use them to decide which interactions to automate.
For a customer service leader, this means that measuring the Customer Satisfaction Score, Net Promoter Score or resolution rate is not enough. These indicators must be linked to individual stages of the customer journey and to the underlying causes of dissatisfaction.
For example, low satisfaction following a chatbot interaction could be caused by:
- An incorrect answer;
- Language that is difficult to understand;
- An excessively long sequence of questions;
- The absence of an option to speak to a human agent quickly;
- A transfer to the contact centre without the relevant context.
An insight becomes genuinely useful when it enables the organisation to modify the customer journey, update the knowledge base, review escalation rules or provide the agent with information already collected by the system.
A Positive Customer Experience Requires a Continuous Cycle
The report shows a significant correlation between the perceived quality of automated experiences and the use of CX insights.
Among organisations that describe their automated experiences as “very positive”, 49% say they continuously optimise automation using customer experience insights. This figure falls to 8% among organisations reporting mixed or negative experiences.
These findings suggest that maturity is not about automating more and more activities, but about establishing a continuous improvement process:
- Collect interactions across all channels;
- Analyse conversations, feedback and operational data;
- Identify friction points, recurring patterns and reasons for contact;
- Decide which activities to automate or redesign;
- Support agents with suggestions and relevant context;
- Measure the outcome of each interaction;
- Use the newly collected data to refine the customer journey.
This approach can be described as intelligence-led automation: automation guided by intelligence derived from customer behaviour, needs and feedback.
The Human Factor Remains Decisive
Artificial intelligence does not eliminate the role of human agents. On the contrary, the report confirms that automation and human expertise work best when designed as complementary capabilities.
As many as 95% of respondents believe human agents add more value than AI in at least one type of interaction. Situations where the human contribution is particularly important include:
- Complex and non-standard issues;
- Complaints and escalations;
- Emotionally sensitive situations;
- Vulnerable customers or those with specific needs;
- Requests with significant consequences for the customer or the organisation;
- Cases requiring negotiation, judgement or empathy.
Furthermore, 73% of organisations report that automated interactions are transferred to a human agent in at least half of cases, considering different types of customer journeys overall. Among organisations with very positive automated experiences, the escalation rate to human agents reaches 76%, compared with 68% among organisations with mixed or negative experiences.
This finding may seem counterintuitive, but it carries an important strategic message: a high escalation rate is not necessarily a sign that automation has failed.
Instead, it may indicate that the organisation has designed the handover between technology and people correctly. The goal is not to prevent customers from speaking to an agent, but to ensure they do not have to struggle to do so.
Effective automation should therefore:
- Recognise its own limitations quickly;
- Transfer customers without losing the context of the interaction;
- Provide agents with the conversation history;
- Explain the reason for the request and the steps already taken;
- Avoid forcing customers to repeat information they have already provided.
AI to Empower Agents
Among organisations using AI, 85% apply it to at least one use case designed to support agents. The most common applications include real-time assistance, productivity improvements, and training or coaching.
Additionally, 99% believe AI improves agents’ ability to perform their roles. The main benefits identified include:
- Resolving customer issues more quickly;
- Receiving guidance and suggestions during conversations;
- Handling more complex or detailed requests;
- Improving the accuracy and consistency of responses;
- Reducing the burden of repetitive tasks.
For customer service teams, this means using AI not merely as a replacement tool, but as a genuine agent copilot.
During a call or chat, the system can, for example:
- Retrieve the most relevant information in real time;
- Suggest the response most consistent with procedures and policies;
- Identify sentiment and urgency levels;
- Flag potential escalation risks;
- Suggest the next step;
- Automatically generate an interaction summary;
- Highlight potential coaching opportunities.
The agent remains responsible for the customer relationship but can work with greater speed, accuracy and continuity.
The Risk: Optimising the Business Instead of the Customer
One of the report’s most significant findings concerns the risk of making automation decisions based primarily on internal objectives.
As many as 76% of organisations agree that automation is implemented to meet internal objectives rather than customer needs.
Reducing costs, managing volumes, improving productivity and shortening handling times are all legitimate goals. However, when they become the only design criteria, they can compromise the quality of the customer experience.
Every automation project should therefore address a number of fundamental questions:
- Does the customer get an answer more quickly?
- Can the customer achieve their goal?
- Is the journey simpler than it was before?
- Does automation reduce or increase the effort required from the customer?
- Is the handover to a human agent simple and transparent?
- Do agents receive the context they need?
- Is the system updated based on real customer conversations?
Effective automation is not the automation that deflects the largest possible number of contacts, but the one that resolves requests correctly and makes the customer journey easier.
How to Turn the Report into Concrete Actions
CallMiner’s findings can be translated into an operational plan for customer service teams:
1. Map the Priority Customer Journeys: identify interactions with the highest volumes, greatest complexity or strongest impact on satisfaction: support requests, complaints, payments, contract changes, cancellations and technical issues.
2. Analyse Real Conversations: do not rely on surveys alone. Call, chat and email transcripts help organisations understand what actually happens during interactions and which obstacles customers encounter.
3. Measure Automation Using CX Metrics: alongside contact deflection, average handling times and costs, monitor:
- Task completion rate;
- Repeated transfers;
- Abandonment rates;
- Repeat contacts;
- Sentiment;
- Customer effort;
- First-contact resolution;
- Satisfaction following escalation.
4. Define Clear Escalation Rules: establish the situations in which an interaction should be transferred to a human agent and ensure that the handover includes the full context of the conversation.
5. Establish Shared Governance: customer service, CX, IT, operations, compliance and risk management teams should work towards common objectives, with clear responsibilities and a structured review process.
6. Use AI to Support Agents Too: AI can improve service quality not only by responding directly to customers, but also by providing agents with assistance, training, suggestions and analytical insights.
The message of the 2026 CX Landscape Report is clear: the next phase of innovation in customer service will not be defined by who automates the most, but by who automates best.
The most mature organisations will be those able to connect three key elements:
- Intelligence extracted from interaction data;
- Automation of processes that are genuinely suited to technology;
- Human expertise in situations requiring empathy, judgement and the ability to solve complex problems.
The goal is not to replace the customer relationship, but to make it simpler, more consistent and more relevant. To achieve this, every interaction must become a source of learning that helps improve the next one.
Vuoi sapere come applicare questi principi alla tua azienda?
Schedule a meeting with an Atombit expert to analyse your customer service processes, identify automation opportunities and define a practical roadmap for integrating AI, data and human expertise into the customer experience.
*Intelligent automation (also known as cognitive automation) is the integration of three main technological components:
- Artificial intelligence (AI);
- Robotic process automation (RPA);
- Business process management (BPM).
This combination makes it possible to automate not only operational tasks but also decision-making processes.
In practice:
- AI analyses structured and unstructured data and generates insights;
- RPA performs repetitive and transactional tasks;
- BPM orchestrates workflows and ensures process consistency.
The result is a system capable of managing end-to-end processes, adapting over time and continuously improving its performance.
Related pages: Customer Experience |