For the past half a decade, the dominant narrative in financial services has been automate, digitise, deploy AI.
It’s what firms want to hear about through their procurement processes, it’s what suppliers want to showcase, and it’s the only thing to be seen talking about on LinkedIn, where (often suspiciously nebulous) revolutions in customer service are happening daily.
Despite inviting a few eye rolls, this is how innovation works, and it’s how we establish the dos, don’ts and best practice within our discipline.
However, moving past social media superlatives and AI noise, the truth is consumers don’t want less human interaction. They want better-timed human interaction.
The firms that are differentiating themselves in our industry are not those that prioritise removing humans from processes, but those that establish when, how, and why a human should become involved in a digital journey.
In short, the real opportunity is this:
- Automation gathers, structures, prepares information, pushes case files through the process and supports during human interaction
- Skilled people engage in the moments that matter
- Firms need to find ways to identify these moments …in the moment.
In regulated environments, the complexity of interactions is unavoidable. Complaints, redress, vulnerability, financial hardship and fraud are not linear, transactional issues that can have a single blueprint for automation or AI applied to them. They are emotionally charged and psychologically layered.
AI is exceptionally good at gathering structured information, pre-populating case files, identifying missing documents, flagging regulatory risks, analysing sentiment signals and predicting likely outcomes. But it is not good at contextual empathy, and in complex processes, context is everything.
The same task, different emotional worlds
Consider two customer scenarios:
- A customer is asked to provide additional information to support a complaint.
- A customer is asked to provide one final document to unlock a redress payment.
Operationally, these requests may look identical: “Please provide X so we can proceed”, but psychologically, they are worlds apart. In scenario one, the request may feel like burden of proof. In scenario two, it feels like the final step before something beneficial. The same process step carries entirely different emotional weight.
Do firms routinely recognise the benefits of human interaction here, or the potential for automated solutions (delivered at the wrong time, or with the wrong tone) to make resolution more challenging?
More haste, less speed
Success is about matching customer energy in real time, ensuring resolution is not achieved at the expense of understanding, that interactions requiring reassurance are not overly automated, and that pace is not seen as the primary success metric, even when it produces the best financial outcomes for the firm.
Sometimes, slowing down for a seven-minute human conversation accelerates trust, cooperation and satisfaction more than multiple emails, and it’s up to firms to understand when to step in, and when to allow customers to summon humans (seamlessly) into the process.
In regulated financial services, this matters profoundly because perceived fairness drives complaint escalation, emotional temperature influences resolution, trust affects long-term retention, and regulatory scrutiny increasingly centres on customer outcomes.
The permutation problem
Every interaction exists within a matrix of variables: emotional state, financial stress level, personal circumstances, trust in the firm, past experience, channel preference, urgency and perceived power imbalance. Even within a single journey type, the permutations are vast.
This is why predicting behaviour and sentiment is so difficult, and why static journey mapping is no longer sufficient. Firms must move from process-led to psychologically aware journey design, as the real challenge isn’t about choosing between AI and humans. It revolves around sequence and timing.
Firms can choose to let AI do the heavy lifting by collecting documentation, triaging cases, identifying vulnerability markers, analysing sentiment trends, pre-drafting communications and highlighting regulatory risks. Skilled, empowered humans can then step in when subjectivity is high, or when needs are specialised. When trained effectively, AI is better at establishing the appropriate action more quickly, while humans are able to execute that action with judgement and empathy when needed.
This helps you engage customers when frustration is rising, when financial vulnerability is present, when requests are burdensome, or when trust needs rebuilding and ambiguity must be avoided.
This isn’t to say that self-serve environments or automated communications aren’t beneficial when working well, but customers need to be able to break out of process stages when doing so will lead to the best outcome.
The risk of over-standardisation in regulated environments
Financial services often leans toward standardisation given the risks associated with customer harm. Consistency protects against regulatory breach. Templates reduce risk. Scripts ensure compliance. But over-standardisation can unintentionally suppress empathy.
Regulators are increasingly focused not just on process compliance, but on customer outcomes and fair treatment. A firm that treats every complaint stage identically may be operationally compliant but experientially tone-deaf.
A more dynamic view of customer journeys
Firms should ask: where are the emotional inflection points? At which moments does reassurance matter more than efficiency? When is automation sufficient, and when is it damaging? How does this journey feel from the customer’s perspective?
Instead of mapping journeys by product or stage, consider mapping them by emotional intensity, perceived fairness, financial consequence, power imbalance and vulnerability risk.
Differentiation over the next decade
Overuse of AI and automation risks outcomes, but indiscriminate human interaction is inefficient. The answer is better human interactions at the right moments. AI should lower the cost of processes, and humans should increase the value of engagement.
As AI becomes ubiquitous, every firm will automate document gathering. Every firm will use predictive analytics. What will differentiate firms is timing, judgement and psychological awareness.
AI and process automation will gather evidence, structure case files, identify comparable precedents, flag regulatory considerations and outline likely outcome ranges before an Agent ever speaks to the customer. This has the benefit of creating ‘ownership moments’ within the customer journey, where Agents have the tools at their disposal to confidently own a customer’s case end-to-end.
In the next decade, the most advanced complaints operations will not be those that remove humans from the process entirely. Instead, those who use automation to ensure they appear at precisely the right moment, informed, empowered and accountable for the outcome.
