Matt Scholes (Global Client Solutions Director) - FinTech and BFSI

Crypto exchanges do not usually lose customers at the moment of trade. They lose them in the moments before, after and around the trade.

A user gets stuck in KYC. A deposit takes longer than expected. A fee is not clearly understood. A market move creates urgency, but support is slow to respond. The customer may still complete the transaction, but the trust has already started to weaken.

That matters because crypto has a retention problem. Studies tracking wallet behaviour suggest that a large proportion of users become inactive within 90 days. Separately, onboarding friction remains a major issue, with many users abandoning before they complete their first transaction. Even allowing for the difference between wallet behaviour, app engagement, and exchange activity, the direction of travel is clear: Getting users to sign up is only half the battle. Getting them to stay is where many platforms struggle.

This is where customer support is often misunderstood.

Many exchanges invest heavily in security, liquidity, product design, and trading functionality, which they absolutely should. But the digital customer experience (CX) around those features is just as important. If users do not understand what is happening, cannot get help when money is involved, or only hear from the platform when something has gone wrong, the relationship becomes fragile.

In a market where trust, speed and confidence matter, customer support should not be treated as a back-office function. It should be part of the operating model that protects retention, customer lifetime value and long-term revenue.

Here are four misconceptions I still see across the sector.

Misconception 1: Treating customer support as a cost center

Too many companies still see customer support mainly as a cost to be contained.

That usually leads to a familiar pattern: Push more users to FAQs, deflect more tickets, reduce human contact and measure success by how quickly cases disappear from the queue. Self-service absolutely has a role to play, especially for simple issues. But when deflection becomes the primary measure of success, support starts being optimised for cost reduction rather than customer retention.

That is risky in crypto because the customer context is different. A delayed response is not just inconvenient. It can involve money, market movement, identity checks, account restrictions, or fraud concerns. Those are high-trust moments.

Put simply, if a platform spends heavily to acquire a user, but then loses that user after one or two transactions because the onboarding or support experience is poor, the economics do not work. The acquisition cost has already been spent. The value is only recovered if the customer stays, trades again and builds confidence in the platform.

This is where customer support becomes a revenue optimization lever. Proactive onboarding, follow-up after first trade, clear escalation paths, reengagement when activity drops and intelligent routing for sensitive cases can all help extend the customer relationship.

The goal is not to throw more people at every problem. The goal is to design customer support around the moments that influence trust, retention and repeat activity.

Misconception 2: VIP programs should only focus on ‘whales’

It makes sense that exchanges give special attention to high-value traders. These customers generate disproportionate volume, expect fast access and are expensive to lose. Dedicated relationship managers, priority support, and tailored service models are logical for that segment.

The mistake is assuming that only today’s highest-value customers deserve a serious support strategy.

Most future high-value users do not look high-value on day one. They might start with a small deposit, test the platform, make a few trades, pause, return, stake assets later, or gradually increase their activity as their confidence grows. If they receive poor support early, they may never reach that point.

That is why a purely top-heavy support model can create long-term risk. It protects today’s largest users, but neglects the broader base that creates future growth, liquidity, advocacy and resilience.

A stronger approach is tiered support. High-value and high-risk customers still receive dedicated service, but everyday users are not left with a thin, generic experience. The right model gives each segment an appropriate level of access, speed and reassurance.

That does not mean treating every customer the same. It means recognising that support is part of how users progress through the relationship. Good CX protects today’s priority accounts while also giving tomorrow’s valuable customers a reason to stay.

Misconception 3: Using AI for CX only to automate tasks

Crypto exchanges already use AI across areas such as fraud detection and prevention, risk monitoring, liquidity management and operational controls. In customer support, however, AI is still often viewed too narrowly: chatbots for FAQs, automated triage, faster case handling and lower cost.

Those use cases matter, but they are only part of the opportunity.

AI for CX can help exchanges understand customer behavior before it becomes churn. Login frequency, transaction patterns, repeated support contacts, failed onboarding steps and changes in activity can all signal when a user is losing confidence or drifting away. Used properly, those signals can trigger smarter interventions before the customer disappears.

Generative AI (GenAI) for CX can also improve how teams use customer intelligence. Voice-of-customer analysis, sentiment tracking, and consumer interaction analytics can help identify recurring pain points across onboarding, payments, account access, fees, staking, fraud alerts and product usability. Instead of treating each ticket as an isolated case, the business can see patterns.

That is where customer support becomes more than a resolution function. It becomes one of the best sources of product, risk and customer intelligence in the organization.

There is also an employee experience angle. AI-powered roleplay simulators can help agents practice difficult scenarios, customer personas and edge cases before they handle live interactions. For complex environments like crypto, where the customer may be anxious, frustrated or financially exposed, that preparation matters.

The best use of AI is not removing humans from the experience entirely. It is helping human teams become faster, more consistent and better informed, especially when the issue requires judgement.
 

Misconception 4: Ramping up CX only when something goes wrong

Crypto does not operate on a neat business-hours cycle. Markets move 24/7. A regulatory headline, a sudden rally, a major liquidation event or a social media post can change customer demand very quickly.

If support capacity only scales after queues have already built up, the damage is often done. Customers do not just remember the outcome. They remember whether the platform was visible, responsive and clear when they needed reassurance.

That is why crisis response should not be the starting point for CX planning. It should be part of a broader operating model that is already designed for volatility.

This includes workforce flexibility, multilingual coverage, clear escalation routes, trust and safety workflows, fraud detection processes, and communication playbooks for moments when customer demand spikes. It also means connecting support with product, marketing, compliance, risk and operations, rather than leaving each function to manage its own part of the customer journey in isolation.

A customer does not experience an exchange in departments. They experience it as one platform. If onboarding, rewards, staking, payments, security and support feel disconnected, confidence drops.

Customer support is more than resolving tickets. It sits across the moments where trust is either strengthened or lost.

CX orchestration as competitive advantage for crypto exchanges

The common mistake across all four misconceptions is treating CX as reactive overhead.

In reality, CX is part of the infrastructure that protects acquisition spend, supports customer lifetime value, improves trust, and helps platforms scale through volatility. Security, liquidity and product features are essential, but they are not enough on their own. Users also need confidence that the platform will support them when something is unclear, urgent or sensitive.

That requires CX orchestration across product, marketing, finance, compliance, trust and safety, and operations. It also requires the right balance of people, process, data and technology.

For crypto exchanges, the opportunity is not simply to have a bigger support team. It is to build a smarter operating model: One that uses automation where it makes sense, human expertise where judgement is needed, analytics where patterns are emerging, and scalable outsourcing where flexibility and execution discipline are critical.

TDCX partners with FinTech and crypto platforms to operationalize this model, delivering end-to-end BPO services and technical support outsourcing, trust and safety solutions for fraud detection and regulatory compliance, CX consulting, and AI- and GenAI-powered solutions that turn customer interactions into measurable business value.