By Debi Potgieter, CCXP, ACXP®, Lifetime CCCX
Human.
Agent.
Representative.
REAL PERSON.
Please.
When a customer starts typing these words in capital letters, the journey has stopped working.
They may have repeated the same information, selected every available option and tried several ways to explain what they need.
By then, they are no longer only trying to solve the original problem.
They are trying to escape the experience.
The customer did not enter an escape room
Escape rooms are great for team building.
People enter them by choice. They expect hidden clues, wrong turns and locked doors. Working out how to escape is part of the fun.
Customer service should not work that way.
A customer should not need to discover the secret phrase that unlocks a human being. They should not have to confuse the chatbot, repeat “agent” five times or become angry enough for the system to recognise that they need help.
The customer’s original problem is already the challenge.
The service journey should not become another one.
An escape room is a wonderful team-building experience.
It is a terrible customer-service design.
Human First does not mean human only
Human First CX does not reject AI.
It means choosing the human outcome before choosing the technology.
AI can be the fastest and easiest way to handle many customer needs. It can help people check balances, track deliveries, reset passwords, update information and find answers without waiting for assistance.
It can also help employees work better. AI can search information, summarise previous conversations and bring useful guidance into view while the employee focuses on the customer.
The problem begins when AI is placed between the customer and help, with no clear way through.
AI can be the front door.
It should never become a locked room.
South Africa is already moving into this shift. A joint report by the Prudential Authority and the Financial Sector Conduct Authority gives the first broad view of AI adoption in South African financial institutions. It found uses across customer interaction, internal processes, risk, fraud detection and cybersecurity. Chatbots are already being used to give customers information about products and services.
The question is therefore not whether AI belongs in customer service.
It is whether we are designing it around the customer outcome, with clear accountability when the technology reaches its limit.
What if there is no door?
A more worrying problem begins when the customer cannot reach a person at all.
The organisation may still employ people somewhere behind the scenes. But from the customer’s point of view, the service has become humanless.
There is no telephone number.
No live-chat option.
No email address.
No way to move beyond the bot.
The customer may type “human”, “agent” and “real person”.
Nothing happens.
They search the website, submit forms and lodge automated appeals, only to receive another automated response.
This is not only customer frustration shared on social media. Ombudsman and regulatory bodies are seeing the same pattern.
Australia’s Telecommunications Industry Ombudsman reviewed 1,537 enquiries about digital platforms received between January 2023 and August 2025. It found that people often approached the Ombudsman after exhausting forums, chatbots and email support without being able to speak to the right person.
The Netherlands Authority for Consumers and Markets has also reported that users of large online platforms are often offered only automated contact points. It warned that this does not meet the requirements of Europe’s Digital Services Act, which protects the right to communicate with platforms through a route that is not entirely automated.
Klarna provides another useful warning. In 2024, the company said its AI assistant was doing work equal to that of 700 full-time agents. By 2025, it had shifted back towards hiring people, with its chief executive acknowledging that too much focus on cost had affected service quality.
The customer is no longer trapped in an escape room with a difficult final clue.
They are in a room with no visible exit.
A human escape route means nothing if there is no human on the other side of the door.
Human First CX therefore requires more than an escalation button. It requires enough human capacity, clear accountability and a real commitment to take over when the technology reaches its limit.
The question is no longer only:
Can the customer find the door?
It is also:
Did we leave anyone on the other side to open it?
Usage is not always preference
In July 2026, Centrica, the owner of British Gas, announced plans to remove about 1,300 roles, including around 500 positions in customer operations.
Its chief executive pointed to changing customer behaviour. He said more than 90% of customers used digital channels first and that customer calls had fallen by 20%.
This was presented as evidence that customers preferred digital service.
That conclusion was challenged. Critics argued that people may use digital channels because speaking to a person has become increasingly difficult. Telephone numbers can be hard to find. Callers face long menus and waiting times. Customers are repeatedly redirected towards self-service.
That distinction matters.
Usage tells us where the customer went.
It does not always tell us why.
An organisation can make digital service the fastest option, the cheapest option or the only clearly visible option. The resulting traffic may then be reported as customer choice.
The customer may use the chatbot because it works brilliantly.
They may also use it because they cannot find another route.
Those are two very different experiences.
Containment is not resolution
Digital containment shows how many interactions were handled without reaching an employee or moving into another channel.
When the customer’s need is genuinely resolved, containment can be good for everyone. The customer receives help quickly, and the organisation avoids unnecessary work.
But containment only tells us where the interaction ended.
It does not tell us whether the problem ended with it.
A customer who does not call may have received the answer they needed.
They may also have given up.
They may have tried another channel, asked someone else for help or quietly decided to leave.
Silence can mean success.
It can also mean surrender.
Organisations therefore need to look beyond chatbot activity, call reduction and containment rates.
Did the customer return with the same problem?
Did they have to repeat the interaction?
Did they understand the answer?
Could they act on it?
Did the solution hold?
These questions show whether customer effort was reduced or simply moved somewhere less visible.
Let the human arrive before the frustration
A 2026 field experiment involving Alibaba’s Taobao customer-service operation offers an important lesson.
The study involved 647 customer-service employees and more than 680,000 conversations. It found that agentic AI shortened customer chats, but customer ratings for AI-handled conversations fell.
Human intervention worked best when the AI reached a technical problem beyond its ability. It was less effective once the customer had already become frustrated. The researchers found that early intervention was important.
By the time a system recognises anger, the customer may already feel ignored.
The employee then inherits two problems: the original need and the frustration created by the journey.
We ask the employee to recover an experience they did not break.
Human First design looks for the moment before that happens.
A handover cannot sit at the end of the journey as a last resort. It should happen when the customer’s situation becomes complex, sensitive, emotional or unusual.
The customer should never have to perform anger to qualify for help.
Put AI beside the human
Bank of America offers a useful example of a different approach.
Its EricaAssist technology works alongside more than 18,000 customer-service employees. During customer conversations, it brings forward relevant information and suggests possible next steps in under three seconds. The bank says it has reduced average call times by almost one minute.
AI carries more of the information load.
The employee carries the relationship.
The technology can find the policy, account information or previous interaction. The employee can listen, understand the situation, use judgement and explain what happens next.
This is where AI can add real value.
It helps the human be more present in the experience.
That is Human First CX in practice: technology supporting people rather than customers being forced to work around the technology.
Design the human route from the start
The route to a person should form part of the original design.
Customers should know when they are dealing with AI. They should be able to reach a person without repeatedly asking. Their conversation history should move with them so that they do not need to start again.
The system should recognise when the situation is becoming complex or emotional. Employees should receive enough information and authority to take ownership.
Customers also deserve honesty about who—or what—is speaking to them.
From 2 August 2026, the European Union’s AI Act requires chatbots and other interactive AI systems to tell users that they are dealing with AI rather than a human. The rule applies in Europe, but the principle is valuable everywhere.
A clear introduction is enough:
I am an AI assistant. I can help you with these requests. For anything outside these areas, I will connect you to a person.
That sets an honest expectation.
Responsible measurement matters too. Resolution, repeat contact, abandonment and customer effort should sit alongside containment.
The central design question is simple:
Are we making the customer prove that their problem deserves human attention?
A clear route to a person does not mean the technology has failed.
It means the experience has been designed responsibly.
Africa cannot afford a locked-room model
This conversation matters deeply in Africa.
By 2024, mobile broadband covered 86% of Africa’s population, yet only 38% of people were using the internet. Internet use reached 57% in urban areas but only 23% in rural communities.
Cost, digital confidence and unreliable connectivity continue to prevent millions of people from fully using digital services.
That matters when organisations treat every move towards a digital channel as customer preference.
A customer may have a mobile signal but limited data.
They may own a smartphone but lack confidence in the journey.
They may reach the chatbot but struggle to explain a complex need in the language or format the system understands.
Digital first cannot be allowed to mean human last.
This also matters because customer service is an important source of employment and growth in South Africa.
The country’s global business services sector created 26,346 new jobs servicing international markets during 2025. Almost 90% of those roles were filled by young people. The sector is working towards a target of 500,000 jobs by 2030.
AI will change this work.
Routine tasks will increasingly be automated. Human roles will need stronger judgement, communication, problem-solving, digital confidence and the ability to work alongside intelligent systems.
We should prepare our people for that future.
At the same time, Africa should not copy digital journeys that hide human help, make customers fight for access or treat vulnerability as an unusual exception.
Our continent is not one market or one customer experience. People use different languages, devices and levels of connectivity. They have different levels of digital confidence and face different financial pressures.
A journey that works for a confident, connected customer with a simple request is not yet an inclusive experience.
The customer who does not understand the instruction matters.
The customer whose circumstances do not fit the available options matters.
The customer facing grief, fraud, financial difficulty, illness or urgency matters.
Africa has the opportunity to build its own model: digital where it makes life easier, human where judgement and care are needed, and honest about the limits of both.
Digital progress should make help easier to reach.
It should never become the reason help is harder to find.
The measure of the digital revolution
The success of digital CX will not be measured by how many customers we keep away from the telephone.
It will be measured by whether customers can move forward with less effort and more trust.
Technology can make service faster, simpler and more accessible. It can remove repetitive work, improve decisions and give employees more time for situations that need judgement and care.
But the final measure remains human.
Was the customer helped?
Did they understand?
Could they act?
Did we protect their time, confidence and dignity?
When automation works, let it work brilliantly.
When it reaches its limit, let humanity arrive early.
The customer should never have to escape the experience we designed to help them.
And where a door is promised, there must be someone on the other side to open it.
Do not make them beg.
Make every interaction count.