The Human Moment

Support That Enables

Engagement

All Sectors

Digital channel design has steadily optimised human contact out of customer journeys. For most low-stakes interactions this is a service improvement: customers prefer a working app to a phone queue. But there are points in every customer journey where the absence of a human is the harm itself — bereavement, claims under stress, financial difficulty disclosure, retirement income decisions, accessibility breakdown. Most firms manage deflection metrics on a portfolio basis, and the numbers improve because the average interaction is well served. The numbers conceal the cohort of interactions that should never have been deflected — the customer at a defining moment who reaches a chatbot, a form, or a hold queue and concludes that the firm cannot help them.
The structural move is to treat human availability as a designed feature of specific journey moments rather than an output of staffing optimisation. The pattern identifies those moments, designs them for human contact, and instruments them so deflection metrics cannot quietly erode them:
Map the human-required moments
Each customer journey has a defined set of moments where stakes, complexity, vulnerability, or emotional load mean that human contact materially changes the outcome. Bereavement notification. Claims under stress. Financial difficulty disclosure. A request to surrender, drawdown, or crystallise. Suspected fraud. A vulnerability disclosure across any channel. Loss of capacity or activation of a power of attorney. The first task is to enumerate these moments per product and journey, evidence why each is human-required, and review the list as journeys change. The design test: can the firm name, for each material customer journey, the moments at which a human must be available, and the evidence base for that judgement?
Engineer human availability into the design
At each named moment, human contact must be designed in — not left to the customer to find. That means proactive routing where signals are detectable (a chatbot keyword, a form abandonment pattern, a search-term cluster), an offered callback when the journey reaches a high-stakes step, a defined handover protocol from digital to human with the customer's context preserved, and a specialist queue with sufficient skilled capacity to handle the moment without rerouting. Where the customer must initiate, the path to a human must be visible and direct — not buried under self-service options that the customer has to bypass. The design test: at each named human moment, what is the path from the customer's first signal to a trained human, and how long does it take?
Measure for harmful deflection, not contact volume
The standard metric set — call volumes, average handling time, chatbot containment, self-service rates — is calibrated against the average interaction and is insensitive to the cases that should not have been deflected. The pattern requires a measurement layer that distinguishes legitimate digital deflection (the customer wanted self-service and got what they needed) from harmful deflection (the customer needed a human and was prevented from reaching one). Channel-mismatch complaints, post-interaction outcome surveys focused on customers in distress, repeat-contact analysis, abandoned-and-not-recontacted cohorts, and outcomes for cohorts crossing the named moments are the relevant data. The design test: does the firm's measurement set reliably surface the human moment that was missed, or does it reward the firm for missing it?
- The firm has a current and reviewed list of human-required moments per product and journey, with evidence supporting each entry and named owners for the design and operational delivery of those moments
- At each named human moment, the customer's path to a trained human is documented end-to-end — including detection signals, routing logic, handover protocol, specialist queue capacity, and target time-to-human — and tested against representative cases
- Outcomes monitoring distinguishes legitimate from harmful deflection: channel-mismatch complaints, repeat-contact rates, abandoned-and-not-recontacted cohorts, and customer-distress indicators are reported alongside contact-volume and containment metrics
- Cross-channel context propagation prevents customers having to re-disclose at the human moment — the bereavement, financial difficulty, or vulnerability disclosure made on one channel is visible to the human handler on the next, without the customer being asked to repeat the information
- A retail bank reviewed customer journeys involving bereavement and power of attorney following the FCA's multi-firm review of vulnerable customer treatment. The review had found common failures: bereaved customers having to repeat sensitive information across multiple staff members, cases dropped or lost in fragmented CRM systems, and limited availability of online or app channels for attorneys, with access constrained by historical investment decisions rather than technical limits. The bank rebuilt the journey around two human-required moments: first contact after a death, and the operational handover when an attorney first acts on the account. A single specialist team was given end-to-end ownership of each case, the customer's context was preserved across all interactions through a unified case record, and the chatbot was reprogrammed to detect bereavement keywords and route immediately to a named handler with the conversation transcript already attached. App-based access was extended to attorneys, removing the differential channel access the review had flagged. Within a year, repeat-information complaints from bereaved customers fell substantially, average days to defund accounts shortened, and senior MI began to track outcome measures — distress reduction, customer satisfaction at completion — alongside SLA performance, which the review had warned could otherwise dominate the picture.
- A wealth platform analysed digital engagement on its drawdown decision tool and found that customers who spent more than three times the average session duration without progressing were significantly more likely to complain about retirement decisions in the months that followed. The platform treated drawdown initiation as a named human-required moment under the pattern. Real-time behavioural monitoring was introduced: when a customer crossed the duration threshold without progression, the journey paused and offered a callback from a retirement specialist within a stated window, with the option to continue digitally if preferred. The flag persisted in the customer record so the next quarterly review, the next withdrawal request, and the next adviser meeting were all aware that the customer had previously hesitated at a defining decision. The platform also reviewed its outcome MI to add channel-mismatch indicators alongside the existing self-service success metrics, so the case where a customer self-served a high-stakes decision under stress would surface in monitoring rather than register as containment success. Complaint rates from the cohort fell, take-up of structured retirement guidance rose, and the FCA's research on digital engagement practices in investment outcomes provided supporting evidence that proactive design at high-stakes moments materially changes outcomes.
Common failure modes
The most common failure mode is treating the pattern as a defence of voice channels in general. It is not. Most journeys benefit from digital optimisation; the pattern is precise about which specific moments must be human, and why. A second is using human availability as compensation for poor digital design — a callback offer when the customer has hit a confusing form is fixing the symptom, not the cause. A third is the channel-mismatch deflection: the customer who reaches the firm via chat about a bereavement, and the bot's natural-language detection misses the disclosure, so the message is queued for a routine response. The FCA's Consumer Support Outcome review specifically called out as good practice the firm whose chatbot automatically routed bereavement queries to a customer support representative — most journeys do not yet do this. A fourth is human-availability that is human-in-name-only: a script-bound agent at an outsourced service provider who cannot deviate from a flow is not a human moment in any operative sense. A fifth is conflating cost with capability — human moments are not unlimited and the Duty does not ask for unlimited human contact, but the firm that cannot evidence which moments warrant the cost has failed the design test that the pattern asks of it.

Related Patterns

The Accessible Default

Reading the Room

The Vulnerability Triage

Evidence & Supporting Facts

FCA Multi-Firm Review of Life Insurers' Bereavement Claim Process (November 2024) found term insurance bereavement claims averaging between 53 and 122 days across firms, with most firms unable to meet some of their own service performance levels and some experiencing extended periods of up to 12 months where SLAs were not consistently met. The review specifically called out support journeys where customers had to repeat sensitive information and cases were dropped or lost.

FCA Retail Banks Multi-Firm Review on Vulnerable Circumstances (April 2025) found a common issue where bereavement and power-of-attorney customers had to repeat information across staff or had cases delayed or lost in fragmented CRM systems. The review observed that limited attorney access to digital channels often reflected investment decisions rather than technical constraints, and that staff failure to recognise distress signals had a direct impact on customer confidence in financial services.

FCA Consumer Support Outcome Review (March 2025) cited as good practice firms that complemented digital channels with human touchpoints whenever complex needs were identified — including a webchat chatbot programmed to route bereavement queries directly to a customer support representative — and noted that 13% of firms surveyed carried out no quality assurance on their support channels. The review specifically warned against reliance on transactional metrics like contact rates and wait times as a substitute for outcome measurement.

FG21/1 (Guidance on the Fair Treatment of Vulnerable Customers) and Chapter 9 of FG22/5 establish that firms must ensure the channels they offer meet the needs of customers with characteristics of vulnerability and customers dealing with non-standard issues, and that flexibility of channel is required for the support to count as Duty-compliant. PRIN 2A.6.2R(3) requires firms to ensure that customer journeys do not contain unreasonable barriers to support.

AI Automation Lens

AI is both threat and opportunity for the human moment. As threat, it accelerates deflection: better chatbots, better IVR routing, better self-service flows can absorb more interactions and make the deflection metric look healthier while concealing the deflected cases that needed a human. The pattern is broken if AI is deployed only against the average. As opportunity, AI is the most powerful detection mechanism available for human-required moments. Natural language processing applied to call transcripts, chat logs, and webchat in real time can identify bereavement, financial distress, suicidal ideation, confusion, and emotional escalation before the customer has explicitly disclosed — the FCA's Retail Banks review noted firms already using speech analytics, in real time and in post-call review, to surface vulnerability indicators and route support accordingly. Behavioural signals on digital journeys — repeated form abandonment, extended session times on a high-stakes decision, multiple failed authentication attempts, drop-off at a settlement step — can trigger a callback offer, a specialist handler, or a held position rather than a closed session. AI can also free human capacity for human moments: automating the routine work that today consumes specialist time means the trained agent is available when bereavement, fraud, or complex claims arrive. The design test is whether the AI investment increases the rate at which the right human reaches the right customer at the right moment — not whether it lowers the contact rate.