Join the waiting list for my next workbook, Strategy in Service Design here. You will receive 20% off when it launches. If you bought one of my previous workbooks and left a review, you’ll receive 50% off instead as a thank you.

If you have been following this series from the beginning, there is a thread connecting everything we have explored so far.

In Volume 1, we looked at service as something much bigger than an interface. In Volume 2, we explored how services behave differently depending on the system around them. In Volume 3, we looked at services as the infrastructure behind modern life. And in Volume 4, we looked at what happens when AI starts participating in the decisions inside that infrastructure.

So, after everything we have explored so far, it feels natural to move from systems, services, and AI into something much more human: how we decide what can still be believed.

How do we know what to trust?

Before trust became something we associated with brands, governments, platforms, institutions, or technology, it began as something much older.

It began with survival.

For early humans, life depended on cooperation. No one could hunt, gather, protect, heal, raise children, share knowledge, or survive danger entirely alone. People had to depend on each other before they had perfect evidence. They had to believe that someone would return with food, keep watch at night, care for a child, share a tool, honour an agreement, or warn the group if danger was nearby.

So trust was not originally an abstract idea. It was a way of living with uncertainty.

This is important because trust has always involved judgement. We decide who to trust based on what we can observe, what others tell us, what has happened before, and whether someone or something has shown enough reliability over time.

In smaller communities, those signals were easier to see. Reputation mattered because people remembered. Behaviour had consequences because people lived close to each other. Trust was personal, social, and visible.

But as societies became larger, trust could no longer depend only on knowing someone directly. It had to become more organised.

Instead of only trusting a person, people started trusting roles, rituals, witnesses, agreements, documents, institutions, and eventually systems.

So the origin of trust matters because it shows us something simple: Trust exists because humans cannot personally verify everything they depend on.

We trust because we have to.

Not blindly, but through signals.

And as services become more complex, more digital, and more automated, the question becomes harder. Because the signals we used to rely on are no longer always visible. We may not know who made the decision, what data shaped it, what system verified it, or where accountability sits.

So the question is no longer only whether people can trust each other.

It is whether they can still trust the systems acting on their behalf.

This is one of the ideas I explore in ”System Thinking in Service Design Workbook”. The workbook helps you look beyond the visible experience and understand the hidden systems, decisions, operations, risks, and responsibilities that shape whether a service can be trusted in the real world.

As we can see, trust began because humans needed to rely on each other before they could verify everything.

But as groups became larger, trust could no longer depend only on direct experience. You could know whether someone close to you was reliable because you had seen their behaviour over time. Once communities expanded, people needed other ways to decide who could be believed.

This is where trust started to become social.

Reputation became a form of evidence. Witnesses helped confirm what happened. Rituals made promises visible. Shared norms created expectations. And social consequences made betrayal more expensive.

Trust was no longer only held between two people. It became something held by the group. This matters because it shows that trust has always needed some kind of infrastructure. Even before banks, courts, passports, contracts, or digital identity, communities were already creating systems to remember behaviour, protect agreements, and decide what should happen when trust was broken.

The infrastructure was not digital. It lived in memory, reputation, witnesses, rituals, rules, and consequences. So trust started as a relationship, but it quickly became a social system.

And once trust becomes a system, it also becomes something that can be designed, protected, manipulated, or broken.

So, following what we just discussed, trust did not stop at reputation.

As societies became bigger, more mobile, and more complex, memory alone was no longer enough. A handshake could work in a small community, where people knew each other and behaviour could be remembered. But it became much harder to rely only on that when people started trading across distance, borrowing money, owning property, crossing borders, studying at institutions, receiving medical care, or dealing with governments.

So, trust had to become more organised.

This is where societies started building systems that could hold trust in place even when people did not know each other personally. Contracts made agreements formal. Courts created a place to resolve disputes. Banks made value transferable. Identity documents made people legible across distance. Professional licensing signalled competence. Universities certified knowledge. Public records created memory outside the individual. Audits, standards, and regulation added another layer of reassurance.

In other words, modern society works because trust has been transferred into systems.

We do not trust a doctor only because they seem convincing. We trust the licence, the training, the institution, and the professional standards behind them. We do not trust money because a stranger says it has value. We trust the banking system, the state, and the infrastructure that supports it. We do not trust a passport because of the paper itself, but because of the institutional authority it represents.

So, little by little, trust stopped being only interpersonal and became infrastructural.

And that matters for everything that comes next, because once trust is embedded into systems, the question is no longer only whether people trust each other.

It becomes whether the systems themselves remain trustworthy.

A useful book to read alongside this section is The Company of Strangers by Paul Seabright. It helps explain how modern society depends on cooperation between people who do not know each other personally, and why trust had to move from direct relationships into markets, institutions, rules, records, and shared systems. For this section, the value of the book is that it shows how everyday life became possible because strangers learned to rely on structures bigger than personal trust.

Once trust moved into organised systems, people needed visible proof that those systems had recognised something as true.

A promise was no longer enough.

People needed evidence.

This is where signatures, stamps, certificates, receipts, photographs, screenshots, and official documents became so powerful. They worked as shortcuts for trust. They allowed people to believe that something had happened, been approved, been paid, been witnessed, or been recognised by a system.

For a long time, evidence helped society move faster.

This is where the idea that “a picture is worth a thousand words” becomes important. Images became powerful because they compressed trust. A photograph could make something feel real. A screenshot could make a transaction feel confirmed. A document could make a decision feel official. A stamp could make authority visible.

But this also created a dependency.

Modern services do not only depend on rules, systems, and institutions.

They depend on evidence being believable.

And that is exactly where AI starts to change the story.

This is also the kind of thinking I explore in ”System Thinking in Service Design“ Workbook. It helps readers understand the hidden evidence, decisions, systems, records, and responsibilities that sit behind a service, and why trust depends on more than what people see on the surface.

As we know, modern services depend on evidence being believable.

A photo, a document, a voice, a video, a receipt, a screenshot, or an identity check only works because people assume there is some connection between the artefact and reality. The evidence does not show the whole system behind it, but it gives people enough confidence to move forward.

This is where AI changes the meaning of proof.

When images, voices, videos, documents, reviews, messages, and even identities can be generated or manipulated, the problem is not only that fake things exist. Fake things have always existed. The deeper problem is scale, realism, and speed. Ofcom found that almost half of UK adults and children believed they had encountered a deepfake at least once in the previous six months, with more than one in ten saying they had encountered them more than ten times.

Fraud is moving in the same direction. The FBI’s 2025 Internet Crime Report said cyber-enabled crimes cost Americans nearly $21 billion, and reported more than 22,000 complaints involving AI-related information. It also warned that scammers are using fake profiles, voice clones, fake documents, and believable videos to impersonate people and institutions.

So AI does not only create fake content.

It weakens the shared signals people use to decide what can be trusted.

And once proof becomes easier to generate, trust can no longer sit only in the artefact itself. It has to move into the system that verifies it, explains it, challenges it, and protects people when the evidence is wrong.

A useful book to read alongside this section is Deepfakes and the Infocalypse by Nina Schick. It is especially relevant because it explores what happens when images, videos, voices, and digital evidence can be generated at scale.

Trust does not break in the same way when the service is optional and when the service is something people depend on to live.

If someone loses trust in a product, they may leave, complain, switch provider, delete the app, or find another tool. That can be frustrating, but there is usually still a way out.

But when trust breaks inside service infrastructure, the person may not have another route.

This is where the issue becomes much bigger than user experience. Around 800 million people globally still do not have an official ID, which means access to banking, work, government support, voting, healthcare, and other services can become difficult or impossible before a person even reaches the service itself. Trust is not abstract here. It is tied to recognition, eligibility, identity, and access.

The same happens with digital services. The UK Government’s Digital Inclusion Action Plan cites Lloyds data showing that only 76% of people with an impairment had foundation-level digital skills, compared with 91% of people without an impairment. It also notes that 30% of people who are offline say the NHS is one of the most difficult organisations to interact with.

So trust is no longer only a brand problem. It is an access problem.

If someone cannot trust how a service verifies identity, makes decisions, explains outcomes, or corrects mistakes, they are not only having a bad experience. They may become powerless inside a system they cannot avoid.

A broken product can lose a customer.

A broken service can trap a person.

I’m currently building my next workbook, Strategy in Service Design, which continues this kind of thinking into trust, evidence, decisions, and accountability in modern services. The waiting list is now open here, and everyone on it will receive 20% off when it launches. If you bought one of my previous workbooks and left a review, you’ll receive 50% off instead as a thank you.

So, here we come into an interesting part because if a broken service can trap a person, then data is not a neutral detail sitting behind the service.

It becomes part of how the service recognises people.

Big data is often presented as objective because it looks mathematical. But data is not produced outside society. It comes from the systems people interact with, the institutions they are exposed to, the records created about them, and the categories available for describing their lives.

And society is not equal.

Some people are constantly recorded, measured, checked, flagged, assessed, and monitored because they have more contact with public services, debt systems, policing, immigration, welfare, healthcare, or housing support. Others are missing from the data because they are offline, undocumented, digitally excluded, informal, mobile, or simply living in ways the system does not recognise.

Globally, the ITU estimates that 2.6 billion people were still offline in 2024, while the World Bank estimates that around 850 million people do not have an official ID. In the UK, the Digital Inclusion Action Plan cites Lloyds data estimating 1.6 million people are offline, and around 23% of the UK population may struggle to interact with online services.

So when services become more data-driven, trust can become unequal.

This is where bias begins to enter the story, but not only as a technical problem inside a model.

Bias can begin much earlier. It can sit in what data exists, whose lives are recorded, which categories are available, what evidence is accepted, and what the service treats as normal.

And this is why trust matters so much.

A data-driven service can only be trusted if it can recognise people fairly, explain how it sees them, and correct itself when the data gets them wrong.

A useful book to read alongside this section is Weapons of Math Destruction by Cathy O’Neil. It helps explain how data-driven systems, scoring models, and automated classifications can appear objective while quietly shaping access to jobs, loans, education, insurance, healthcare, and opportunity. For this section, the value of the book is that it shows how inequality can become embedded into systems when data is treated as neutral, even when the society producing that data is not.

If trust becomes unequal when data sees people unequally, the next question is practical.

What do we actually do with this as service designers?

I think this is where the discipline becomes much more important in the next few years. Not because service designers will “own trust” as a concept, but because trust lives in the exact places service design is supposed to understand: decisions, evidence, operations, escalation, exceptions, responsibility, and the way people move through systems over time.

Trust is not something people simply feel because the interface looks clean. Trust is something a service earns through how it behaves.

A trustworthy service shows people what is happening, why it is happening, who is responsible, what evidence is being used, what options exist, and what can be done when the system gets something wrong.

So, when designing future services, I would pay attention to:

This is also where future careers in service design will evolve.

The most valuable people will not only be the ones who can map a journey or facilitate a workshop. They will be the ones who can understand how trust is produced across a service system. They will know how policy becomes rules, how rules become decisions, how decisions become outcomes, and how those outcomes affect real people.

This means learning more about governance, risk, evidence, AI, operations, data, accountability, accessibility, and institutional behaviour. Not to become an expert in everything.

But to become better at seeing what other disciplines miss when they only look at their own layer.

Because the future of service design is not only about improving experiences. It is about helping organisations build systems people can understand, challenge, and rely on.

Everyone on the waiting list will receive 20% off during the first 48 hours of launch. If you bought System Thinking in Service Design and left a review, you’ll receive 50% off my new workbook as a thank you.

If you’d like to explore more topics around service, systems, AI, and organisational complexity, you can also follow me on Linkedin.

This volume was about understanding trust as part of the infrastructure of modern services.

We started with trust at the human level, then followed how it moved into reputation, institutions, evidence, records, systems, and now digital services. As AI makes proof easier to generate and data systems become more influential, the question is no longer only whether people trust what they see. It is whether services can still prove, explain, correct, and take responsibility for the decisions they make.

Next, we move from trust to bias.

If trust depends on how services recognise people, then bias becomes one of the biggest risks inside AI-driven systems. The next volume will explore how inequality enters services through data, categories, policies, assumptions, historical patterns, and institutional behaviour.

Because AI does not learn society from a neutral place. It learns from the systems we have already built. And when those systems are unequal, AI can turn old patterns of exclusion into faster, larger, and more invisible service outcomes.