Designing Digital Trust at Scale With Oliver From Engine

A digital journey can complete successfully while leaving the customer uncertain about what happened, why a decision was made, or what they should do next.

That uncertainty can surface later through repeat contact, abandonment, complaints, and reluctance to trust the next interaction.

In this episode of Regulated Digital, we speak with Oliver King, co-founder and managing partner at Engine, about how regulated organisations can design confidence into digital journeys as automation and AI take on more of the process.

Our questions centre around where trust appears inside a journey. It may be a clear explanation before an identity check, a useful progress update during a delay, or an obvious route to a person when a decision has serious consequences - but these details shape whether the customer feels safe enough to continue.

Engine can help you identify which customer problem is destroying the most value in your business with a CX Trust Audit®.

Created from episode transcript

Trust Is Felt as Confidence in the Next Step

Trust is difficult to manage when it remains an abstract brand value. Oliver's more useful test is to look at how confident the customer feels at each moment. Do they understand what is happening, believe the organisation is in control, and know what to do next?

Confidence builds through small signals: consistent answers, clear ownership, relevant explanations, realistic timeframes, and visible routes to help. It drains when customers repeat information, receive conflicting guidance, wait without knowing why, or open a message that creates more questions than it answers.

"Confidence isn't created by one big moment. It comes through hundreds of tiny interactions. Every message, handover, and explanation either adds a little confidence or, if handled badly, quietly strips it away."

Oliver King, Co-founder and Managing Partner, Engine

This changes the job of communication. Sending an update only helps when it reduces uncertainty. A message that confirms receipt, names the next action, gives a realistic timeframe, and explains what the customer can do in the meantime performs a different role from a generic status notification.

Efficiency Can Remove Useful Reassurance

Digital teams are often measured on speed, completion, conversion, and cost to serve. Those measures are useful, but they can encourage teams to remove steps without checking what each step contributed to the customer's understanding or sense of control.

The FCA's review of digital customer journeys warns that design choices can drive customers towards quick decisions and that firms should consider whether a journey has an appropriate amount of friction. It also advises firms to use multiple indicators when assessing outcomes, because positive online reviews alone are unlikely to represent the whole customer base.

Completion rate is particularly easy to misread. A customer may finish a mandatory claim, transfer, or account-recovery process because there is no alternative. The dashboard records success even when the person had to leave and return, search for help, switch channels, or call afterwards for reassurance.

The deeper failure appears when a journey becomes operationally efficient but emotionally uncertain. The organisation saves seconds or removes a contact, while the customer delays a decision, makes another enquiry, or becomes less willing to trust the next automated interaction.

Good Friction Creates Time, Understanding, and Control

Some friction is plainly wasteful. Asking for the same information twice, forcing customers through multiple logins, hiding important terms, or making them wait without explanation adds effort while reducing confidence.

Useful friction gives the customer something in return. It can create time to review a high-value payment, explain why identity verification is required, highlight the consequence of cancelling a product, or confirm that an important instruction is correct before it is submitted.

"Good friction slows someone down just enough to increase confidence. It gives them the opportunity to review an important decision before they commit."

Oliver King, Co-founder and Managing Partner, Engine

This distinction is reflected in the FCA's Consumer Duty guidance. Firms should include appropriate friction where it mitigates the risk of harm and gives customers time to understand and assess their options, while avoiding unreasonable barriers during the product or service lifecycle.

Oliver suggests a simple workshop test: after this interaction, is the customer likely to feel more or less confident? That question keeps attention on the effect of the step rather than the number of clicks it adds.

AI Needs to Explain Decisions and Leave a Route to Act

As AI handles more analysis, triage, recommendations, and routine service work, the decisive customer moment may arrive after the automated processing has finished. A mortgage applicant, claimant, or patient still needs to understand the outcome, the factors that influenced it, and the options available next.

Although written for public-sector organisations, the UK Government AI Playbook provides a useful design standard. It links transparency and explainability with accountability and the ability to appeal, and recommends making people aware when AI is used, explaining which factors influence decisions where possible, and signposting routes for challenge or redress.

The explanation is part of the service. A refusal with no context can make the organisation feel arbitrary. A clear explanation of the decision, what information shaped it, what the person can do next, and how to reach a human gives them a practical way forward.

"In secure customer communication, confidence comes from what the recipient can recognise and do at each step. They need to know who is contacting them, why access is being checked, and what will happen after they respond. Those details can reduce uncertainty while keeping the right safeguards in place."

Paul Holland, Founder and CEO, Beyond Encryption (Mailock)

Human support should therefore be designed around moments where judgement, reassurance, or accountability carries particular weight. That may be after an adverse decision, during a long wait, when responsibility moves between teams, or when the customer signals uncertainty by repeatedly seeking help.

Technology can manage more of the process while the organisation remains accountable for the outcome. The customer should still know who is responsible, how to challenge a decision, and where to ask for help, even when no employee is present in the interaction itself.

Measure Where Confidence Rises and Drains

Trust measurement becomes more useful when teams stop treating it as a single survey score. NPS, CSAT, complaints, and completion rates describe part of the picture, usually after the event. They rarely identify the exact point where uncertainty began.

The FCA's 2026 consumer understanding review recommends combining evidence from call listening, complaints, chat transcripts, website analytics, drop-off data, surveys, and customer testing. This creates a stronger basis for finding where people struggle and checking whether changes genuinely improve understanding.

Oliver describes using emotional analytics to examine large volumes of written feedback and identify the feelings behind what customers say. Combined with journey data, this can help teams connect specific moments of uncertainty with observable behaviour, such as hesitation, abandonment, repeat contact, or channel switching.

This approach gives leaders a clearer commercial case for change. Teams can show where a confidence drain increases support demand, delays a decision, or weakens the likelihood of a customer returning.

Start With One Journey and a 90-Day Test

A full trust programme can become too broad to act on. Oliver recommends beginning with one important journey and changing the question from "How can we make this faster?" to "Where does confidence rise and fall?"

First, review existing evidence to locate the largest confidence drains. Complaints, call reasons, handovers, repeated questions, abandoned forms, and help-page visits can reveal where customers stop feeling informed or in control.

Next, redesign a small number of moments. Add a clearer explanation, name the team responsible, set an honest timeframe, show the next step, or offer a route to a person at the point where uncertainty is highest. These changes often require better content and ownership rather than a new platform.

Finally, test whether the changes alter both behaviour and understanding. The FCA's consumer understanding guidance recommends testing before and after changes, using methods such as comprehension checks, surveys, A/B testing, and frontline feedback. The aim is evidence that customers understood and could act, rather than a cleaner-looking screen alone.

As the technology used across regulated services becomes more similar, these accumulated confidence signals can become a source of distinctiveness. Competitors can adopt the same models and interfaces. The pattern of explanations, handovers, safeguards, and human support across the whole relationship is harder to reproduce.

FAQs

What Is a Confidence Signal in a Digital Journey?

A confidence signal is a small cue that helps a customer understand what is happening and feel safe continuing. Examples include a clear explanation, visible ownership, a realistic timeframe, confirmation of the next step, or an obvious route to help.

How Can Firms Distinguish Good Friction From Bad Friction?

Assess what the step gives the customer. Good friction supports understanding, control, security, or a considered decision. Bad friction adds effort, delay, or confusion without a clear customer benefit.

Which Metrics Can Show Whether Trust Is Rising or Falling?

Use operational, behavioural, and emotional evidence together. Completion and complaints are useful, but they become more informative when combined with hesitation, backtracking, repeat contact, channel switching, and the language customers use in feedback.

Where Should Humans Remain in AI-Led Journeys?

Prioritise moments involving adverse decisions, vulnerability, complex judgement, long waits, disputed outcomes, or signs that the customer is uncertain. The route to a person should be designed into the journey rather than added only after failure.

References

Digital design in customers' online journeys: good practice and areas for improvement, Financial Conduct Authority, 2025

Consumer understanding: good practice and areas for improvement, Financial Conduct Authority, 2026

FG22/5 Final non-Handbook Guidance for firms on the Consumer Duty, Financial Conduct Authority, 2022

Artificial Intelligence Playbook for the UK Government, Government Digital Service, 2025

Reviewed by

Sam Kendall, 14.07.26