A VP of revenue operations at a mid-size consumer lender pulled a report last Tuesday that changed her quarterly plan. Of the 62,000 customers who contacted support in the previous month, 15,000 were making payments on a personal loan while carrying credit card balances above $8,000. Every one of them was a candidate for a debt consolidation product the company already offered. Not one had been offered it. The customers called about payment dates, asked about interest calculations, and hung up. The revenue opportunity hung up with them.

That lender is not unusual. Salesforce's 2024 State of Service report found that 85% of service decision makers say their teams are expected to contribute a larger share of revenue through cross-selling, upselling, and retention in the coming year. The expectation is there. The execution, in most organizations, is not.

The gap is especially wide in regulated industries. Banks, lenders, insurers, and healthcare companies sit on enormous cross-sell potential inside their support queues. But compliance requirements and the risk of regulatory action make most of them hesitant to try. The revenue stays on the table, and the customers who would benefit from a better-fit product never hear about it.

The revenue hiding in support.

Customer support interactions are the highest-intent sales moments most companies ignore. A customer who calls about their personal loan payment is already engaged with their financial relationship. They are thinking about money. They are on the line. They have context. That is a fundamentally different selling environment than a cold email or a banner ad.

The data confirms this. The probability of selling to an existing customer is 60% to 70%, compared to just 5% to 20% for a new prospect. McKinsey's research on contact center revenue found that companies focusing on service-to-sales conversion saw a 40% increase in sales conversion rates from service calls, while also improving customer satisfaction and stabilizing handle times.

Cross-selling already accounts for up to 30% of revenue for many companies. Amazon attributes 35% of its revenue to cross-sell recommendations. Yet most financial services companies treat their support channel as pure cost, measuring success by how fast they can end the call rather than how much value they can create during it.

The math on the lender above is straightforward. If even 10% of those 15,000 monthly candidates convert on a consolidation loan averaging $12,000, that is $18 million in new loan originations per month from a channel that currently generates zero. A pilot program at that lender converted at 18%, suggesting the actual number is considerably higher.

What is Lorikeet?

Lorikeet is an AI customer experience platform purpose-built for complex and regulated businesses. It resolves customer interactions end-to-end across chat, email, and voice, handling account inquiries, payment processing, claims, billing adjustments, and multi-step workflows that require pulling data from back-end systems and taking action on behalf of the customer. Lorikeet operates within defined compliance boundaries, ensuring that every customer interaction, including cross-sell recommendations, adheres to regulatory requirements and approved product guidelines.

Why support beats cold outreach.

The timing advantage of cross-selling during a support interaction is difficult to overstate. Harvard Business Review research found that difficult service interactions have only a 6% chance of resulting in a cross-sell, while excellent service experiences produce an 80% chance. The quality of the support interaction is the single biggest predictor of whether a customer is receptive to hearing about another product.

When AI resolves a customer's issue quickly and accurately, the trust generated in that moment opens a window for a relevant recommendation. A marketing email about a savings account gets ignored 97% of the time. The same customer, ten seconds after an AI agent confirmed their next payment date, is in an entirely different mental state. They are satisfied. They trust the interaction. A mention that their payment history qualifies them for a lower-rate credit line lands differently than a blast email ever could.

USAA has built its cross-sell approach around exactly this principle. It mines customer data to identify life events and contacts customers at the right moment with the right offer, turning service interactions into the primary distribution channel for new products.

The compliance problem.

In unregulated industries, cross-selling during support is simply a matter of training and tooling. In financial services, healthcare, and insurance, it is a minefield.

The Wells Fargo scandal remains the cautionary tale. Employees opened more than two million unauthorized accounts to hit cross-sell targets. The CFPB levied a $100 million fine. In 2022, an additional consent order required more than $2 billion in customer redress and a $1.7 billion civil penalty. The company spent nearly a decade unwinding consent orders.

That history has made the entire financial services industry allergic to the word "cross-sell." But the problem at Wells Fargo was not cross-selling itself. It was unauthorized cross-selling, driven by misaligned incentives, with no compliance controls and no customer consent. The answer is not to abandon cross-selling. It is to do it correctly, with full transparency, genuine customer benefit, and verifiable compliance at every step.

The OCC and CFPB have made clear that the regulatory concern is not product recommendations themselves, but deceptive marketing practices, lack of customer consent documentation, inadequate monitoring, and failure to ensure that cross-sold products deliver genuine value. A compliant cross-sell recommendation that helps a customer consolidate high-interest debt into a lower-rate loan is a consumer benefit. Regulators want that to happen. They want it to happen honestly.

How AI changes the equation.

Human agents attempting cross-sell during support face three simultaneous problems: resolve the primary issue, identify whether a cross-sell opportunity exists, verify eligibility, and deliver the recommendation in compliant language. All while maintaining call quality. Most agents skip the cross-sell.

AI eliminates the cognitive load problem. While resolving the customer's issue, an AI agent in a fintech context can simultaneously check the customer's product portfolio, identify gaps, verify eligibility against underwriting criteria, and determine whether a recommendation is appropriate given the customer's current interaction sentiment and history.

The recommendation happens after the primary issue is resolved, not during it. The AI answers the payment question, confirms resolution, and then says: "Based on your account history, you may qualify for a consolidation rate that could reduce your monthly payments. Would you like to hear more?" The customer says yes or no.

AI-powered recommendations have increased cross-selling success rates by 35% in documented implementations. The improvement comes from three factors: better targeting (the AI identifies genuine fit, not random offers), better timing (after resolution, not during the problem), and better compliance (every recommendation follows approved scripts and eligibility rules).

Building compliant cross-sell.

A compliant AI cross-sell system in a regulated industry requires five layers working together.

Eligibility verification. Before any recommendation, the AI checks the customer's credit profile, existing products, account standing, and regulatory restrictions. A customer in a hardship program does not receive a credit offer. The check happens in milliseconds using the same back-end systems the support platform already connects to.

Suitability logic. Not every eligible customer should receive every offer. A customer calling to report a missed payment is not the right candidate for a new credit product. The AI evaluates interaction context, including sentiment and reason for contact, to determine whether a recommendation is appropriate in this specific moment.

Approved language. Every cross-sell message uses language reviewed and approved by compliance and legal teams. The AI does not generate ad-hoc product descriptions or make unapproved claims about rates or benefits. This is where generic AI tools fail in regulated industries. A general-purpose chatbot cannot guarantee that a product description meets Truth in Lending Act disclosure requirements.

Consent documentation. Every interaction is logged with full audit trail: what was offered, the customer's response, and the eligibility data that supported the recommendation. If a regulator asks why a specific customer received a specific offer, the answer is retrievable in seconds.

Opt-out controls. Customers who decline are flagged to prevent repeated offers. Customers who opt out entirely are respected across every channel. The system treats "no" as a durable preference, not a temporary objection to overcome.

Every cross-sell recommendation should make the customer's financial situation better, not just the lender's revenue line better. When that principle is built into the system architecture, compliance becomes a feature rather than a constraint. See how Lorikeet builds compliant cross-sell into support workflows.

Measuring what matters.

The metrics for support-driven cross-sell are different from traditional sales metrics because the primary objective of the interaction is still service. Revenue is the byproduct of good service, not the replacement for it.

Offer rate. What percentage of support interactions include a cross-sell recommendation? Most service-to-sales programs find the sweet spot between 15% and 30%. Below that, the eligibility criteria are too restrictive. Above it, the system risks annoying customers.

Acceptance rate. Of customers who receive a recommendation, how many express interest? This measures relevance. McKinsey found that the difference in sales-conversion rate between top and bottom performers in the same contact center was 230%, driven almost entirely by how well the offer matched the customer's actual need.

Conversion rate. Of customers who express interest, how many complete the application or purchase? This measures follow-through. In the consumer lending pilot described at the top of this article, the 18% conversion rate from offer to completed application exceeded every outbound marketing channel the company operated.

CSAT impact. This is the metric that determines whether the program survives. If cross-sell recommendations reduce customer satisfaction scores, the program is broken regardless of its revenue contribution. The research is clear: when done after successful resolution, cross-sell does not harm satisfaction and can actually improve it, because customers perceive relevant recommendations as a sign that the company understands their needs.

Revenue per interaction. The metric that turns the CFO into an advocate. When support interactions generate measurable revenue, the entire funding model for the service organization shifts. Tracking the right support metrics means the contact center stops being a cost line and starts being a revenue channel with its own P&L.

The data advantage.

Financial services companies already have the data they need to identify cross-sell candidates. It sits in core banking systems, CRM platforms, and loan origination software. The problem is that support systems have historically been disconnected from product and eligibility systems.

AI closes that gap in real time. When a customer calls about their auto loan payment, the AI can simultaneously see their credit card utilization and identify that they are paying 22% APR on a balance that could be consolidated into a 9% personal loan. That identification happens during the interaction, not in a batch marketing campaign three weeks later.

High-performing financial institutions added an average of 1.34 new products per digital banking user in 2024, 24% higher than mid-tier performers and double the rate of low performers. The difference is the ability to identify and act on cross-sell moments in real time.

Lorikeet in practice.

Lorikeet's architecture is built for exactly this use case. Because the platform connects to back-end systems to resolve issues, those same connections enable real-time eligibility checks and product matching during any customer interaction.

When a consumer lending customer contacts Lorikeet about a payment question, the platform resolves the inquiry first. Once the issue is handled, Lorikeet checks their product portfolio against the lender's cross-sell rules. If there is a match and the interaction context is appropriate, it presents a compliant recommendation using pre-approved language and routes interested customers to the appropriate next step.

Every recommendation is logged. Every eligibility check is documented. The compliance team can audit any cross-sell interaction from any channel at any time.

For a VP of RevOps like Diana, who discovered 15,000 cross-sell candidates per month flowing through support unremarked, Lorikeet turns that discovery into a repeatable, measurable, and auditable revenue program. The 18% conversion rate from her pilot becomes the baseline, not the ceiling, because AI-driven targeting improves with every interaction as the system learns which customers, products, and moments produce the best outcomes.

From cost center to revenue channel.

The shift from cost center to revenue channel is already underway. Salesforce found that 82% of high-performing organizations now use the same CRM platform across service, sales, and marketing, up from 62% two years ago. The operational barriers between service and sales are dissolving.

In regulated industries, the question is not whether to cross-sell during support. It is whether to do it with the compliance controls and customer-first design that regulators demand. Bain & Company's research confirms the underlying economics: increasing customer retention by just 5% can increase profits by 25% to 95%. Cross-selling during support deepens the customer relationship, increases product stickiness, and makes the customer harder to lose. A customer with one product is a transaction. A customer with three products is a relationship.

The 15,000 monthly candidates in Diana's support queue are not an anomaly. They are the norm at every mid-size and large financial services company. Lorikeet makes it possible to pay attention at scale, with compliance built in, and turn every support interaction into a chance to make the customer's financial life better while making the business stronger.