A dissatisfied customer will complain and, at most, won’t do business with a service or use the service again. When a customer is angry, they typically take other types of action: discourage others from using the product or service, or go to court against the company. Emotion has a significant impact on customer experience and customer perception of value related to the product. Emotionally tuned customer experiences can significantly boost business metrics. For example, a bank’s emotionally designed credit card saw a 70% usage increase among millennials and a 40% account growth (Scott Magids, 2015).
This article explores how Agile leaders can enhance customer satisfaction by integrating emotion recognition and Emotional AI into product and service delivery. Learn how to use classes of service, AI-driven insights, and human empathy to adapt faster, prioritise better, and deliver greater value in today’s emotion-driven marketplace.
Who is this article for?
“Food for Thought” is a series of mini-articles containing inspiring content designed to enrich the perspectives of conscious, modern agile leaders. Each article in this series aims to trigger reflection, inspire further exploration, and serve as a starting point for fruitful discussions.
The Food for Thought series is dedicated to thoughtful agile leaders in different forms:
- Senior Leaders, Team Leaders, Product Leaders, People Leaders and Organisational Leaders interested in maturing their leadership
- Product Owners, Scrum Masters, Agile Coaches, and Delivery Managers interested in developing their leadership.
- All those who see themselves as modern Agile leaders who continuously seek new ways to develop their craft.
In the following paragraphs, I will briefly explain emotional detection, then explain the concept of classes of service and how the classes of service are defined. Afterwards, I will detail how classes of service can be enriched with customer emotions. I will also briefly explain how the concept can be implemented, along with examples of emotional AI tools that can be used for service and product delivery. The article concludes with a final summary.
Emotional detection
Emotion recognition is the process of identifying the emotions of people around you. Being able to detect, process, and respond appropriately to the emotions of others is crucial for healthy and effective social interaction (Ben Corden, 2006). Interpreting another’s emotion provides insight into their thoughts, beliefs, and intentions and allows one to explain and interpret their behaviour and, to some extent, interpret their future actions
People express emotions through various channels, often combining several at once to convey their feelings. The main modes include:
- Facial Expressions -Universally recognised indicators like smiling, frowning, or crying reflect basic emotions (Ekman, 1992).
- Body Language and Posture – Gestures, posture, and movement often reveal emotional states unconsciously. (Argyle, 2013)
- Vocal Tone and Speech Patterns – How something is said (tone, pitch, volume) conveys emotions beyond words (Scherer, 2003).
- Verbal Expression – Direct communication using emotional language to describe feelings (Pennebaker, 1997).
- Behaviour and Actions – Emotions can drive actions such as slamming a door or hugging someone.
- Physiological Reactions – Blushing, sweating, or changes in heart rate can visibly express emotional arousal.
Classes of Service, customisation for product and service delivery
Classes of Service are one of the most overlooked concepts in Kanban. When deployed correctly, they can significantly simplify prioritisation of work, improve the risk management process, and impact customer satisfaction. Classes of service are special policies that define how a subset of the work or demand for future work will be treated in the delivery system. By assigning work to these classes, teams can effectively manage the demand for the service in a way that meets customer expectations. In terms of building products, we can use classes of service for product discovery (to learn more, read this article) and delivery.
Typically, team prototype classes of service are based on archetypes of the cost of delays. Cost of Delay Archetypes refer to patterns categorising the economic impact of delaying work. They help prioritise tasks by quantifying the cost of postponing delivery. Common archetypes include:
- Urgent (Expedite): High, immediate cost if delayed (e.g., critical bugs).
- Fixed Date: Cost spikes after a deadline (e.g., regulatory compliance).
- Standard: Linear cost increase over time (e.g., typical features).
- Intangible: Low initial cost but growing long-term impact (e.g., technical debt)
The picture above shows archetypes of the cost of delay. A graph presents the relationship between the time associated with delaying the work and the cost of not delivering specific work.
Using classes of service to adapt delivery to customer emotional needs.The cost of delays is a commonly used factor to define classes of service, but it does not need to be the only factor. (Kelly M. Wilder, 2014) defines the two major processes in adapting a service experience based on customer emotions:
Recognition of customer needs through employee (person facing customer) empathy or usage of a support tool,
- Alternatives in how customer requests are handled and delivered to meet the needs arising from the recognition step.
These two major processes can be effectively associated with classes of service. We can incorporate customer emotion into the process:
- Initial qualification policies define when a specific work item or customer request can be qualified for a particular class.
- Revision of the assignment CoS. Any interaction with a customer is a great trigger to evaluate the current assignment of CoS; something might impact the current customer’s emotions.
- Police associated with specific CoS may include needs related to customer emotions, i.e., if the customer feels anger, we may improve the service level for this particular request.
The schematic model of associating customer emotions with classes of service is presented below.
The schematic flow of adapting classes of service based on customer emotions is presented above. Customer emotion input is presented on the top path. The process of recognition of customer emotion can be supported by dedicated Emotional AI tools, which are briefly described in the next paragraph.
The initial assignment of classes of service can be adapted whenever a new customer emotion is discovered (typically, we find out about them during new interaction with the customer through various channels described in the above paragraph) or whenever new information impacting non-emotional assessment is discovered.
In that way, adapting classes of service based on customer emotions is an ongoing process triggered by discovering new factors affecting the product delivery. Classes of service have predefined sets of polices that help in addressing specific needs of customers (including adapting delivery to better respond to customers’ emotions).
Emotional AI Tools for detecting customer emotions
Previously, from a business perspective, customer emotions and their impact were primarily the concern of employees and their emotional intelligence, which is yet another reason why Emotional Intelligence is one of the most crucial soft skills for a Product Owner.
This situation will not change soon. However, in the short term, humans will acquire powerful tools that will simplify and reduce the cost of recognising and synthesising customer emotions on a large scale, like never before. This change will be triggered by the development of Emotional AI, also known as affective computing.
Emotional artificial intelligence systems are those that can recognise, interpret, and in some cases respond to human emotions. This technology combines data from various channels of human emotional expression, such as voice tone, body language, and physiological signals, to assess emotional states in real time. We can observe an increase in the use of these systems in customer service, healthcare, education, and automotive products and services. Some examples of already existing software can be helpful in emotion recognition. A few of the emotional AI tools are listed and briefly described in the table below.
Text version:
Affectiva
Facial, Voice
Detects emotions from facial expressions and voice, applied in automotive, media, and healthcare.
Viso Suite
Facial
No-code platform for real-time facial emotion recognition, customizable for marketing and education.
MorphCast
Facial
Browser-based facial emotion recognition for virtual meetings and engagement analytics.
Cogito
Voice
Real-time voice emotion analysis for call centers and mental health applications.
Realeyes
Facial
Webcam-based emotion analysis for marketing, optimizing ad content based on audience reactions.
iMotions
Multimodal
Combines facial coding, eye tracking, and biometrics for emotion research in marketing and UX.
Lettria
Text
NLP-based API for emotion extraction from text, used in market research and customer feedback.
NLP Cloud
Text
Emotion detection API for text analysis, integrable into chatbots and sentiment platforms
Behavioral Signals
Voice
Voice-based emotion recognition for call centers, analyzing vocal cues like pitch and tone.
Brand24
Text
Social listening tool using NLP to detect emotions in brand mentions for marketing insights.
Emotional AI: The future of service and product delivery
Teams incorporating emotion into service and product delivery will gain a significant competitive advantage in the competitive market. Current experience in the field of product and service delivery is increasingly digitalised, and both of them are dedicated to addressing the needs of human beings. Humans are emotional and do not make all their decisions based on logic. Emotions are central in shaping how people make decisions, assess value related to a product or perceive service quality. The “risk-as-feelings” theory argues that people’s emotional reactions to risky situations can override logical assessments. These instinctive feelings often guide decisions more strongly than cognitive evaluations (Weber, 2001). This article presents a fresh perspective on how classes of service, at least partially driven by emotion recognition supported with the emotional AI tools, can help better meet customer expectations.
For executives, senior leadership, and teams seeking to accelerate organisational transformation, drive leadership development, or receive support in their digital, business, operational, or Agile transformation, visit pawelrola.com or contact me on LinkedIn.
Let’s work together to bring the organisations of the future!
This article was originally published at: pawelrola.com Direct Lin
Bibliography
Scott Magids, A. Z. (2015). The New Science of Customer Emotions: A better way to drive growth and profitability. https://hbr.org/.
Kelly M. Wilder, J. E. (2014). Tailoring to Customers’ Needs: Understanding How to Promote an Adaptive Service Experience With Frontline Employees. Journal of Service Research.
Ben Corden, H. D. (2006). Fear Recognition Ability Predicts Differences in Social Cognitive and Neural Functioning in Men. Journal of Cognitive Neuroscience, 889–897.
Ekman, P. (1992). An argument for basic emotions. Cognition & Emotion, 169–200.
Argyle, M. (2013). Bodily Communication. London: Routledge.
Scherer, K. R. (2003). Vocal communication of emotion: A review of research paradigms. Speech Communication, 227–256.
Pennebaker, J. W. (1997). Writing about Emotional Experiences as a Therapeutic Process. Psychological Science, 162-166.
Weber, E. &. (2001). Risk As Feelings. . Psychological bulletin 127 , 267-86.