Imagine your company had an AI-powered engine that could detect when a customer needs help—before they even realize it. And was able to coordinate and sequence customer touchpoints while personalizing communications to get the right message to the customer in the right way.
Through our work with leading companies, we see that some are already using this approach—called next best experience—to impressive effect. With properly-calibrated AI models accessing integrated data sets that span the entire customer life cycle, companies can craft experiences that deliver real value.
The results are impressive: The AI-powered next best experience capability can enhance customer satisfaction by 15 to 20 percent, increase revenue by 5 to 8 percent, and reduce the cost to serve by 20 to 30 percent.
This article explores how the AI-powered next best experience approach addresses critical customer experience issues, details what it takes to build and operate it, highlights case studies that demonstrate its value, and outlines six practical steps for organizations to get started.
What is next best experience?
AI-powered next best experience is a customer experience capability that allows companies to proactively deliver each customer the right interaction at the right time in the right place. This approach differs from the more-common “push” approach, where a company essentially spams customers with offers or promotions.
How next best experience works: A journey to satisfaction and delight
In this illustrative example, Thomas is a health plan customer who has diabetes. Before his insurer adopts a next best experience approach, Thomas’s interactions with his medical insurance provider were frustrating.
On one occasion, he incorrectly filed a claim and, when charged a fee for the error, he contacted the customer care line to complain. He had to make multiple calls to sort out the issue—increasing both his frustration and his insurer’s cost to serve.
At the same time, he received an invitation to join a diabetes management program, which he declined due to his recent negative experience. Thomas then forgot to pay his health plan bill, leading to missed payments and a late charge. To top it all, he received a customer experience survey from the insurer, which he did not complete due to his dissatisfaction with his experience (exhibit).
When his insurer implements a next best experience model, Thomas’s experience journey improves. The company moves from a series of uncoordinated interactions to a set of coordinated interventions that focus on improving customer lifetime value (CLTV).1
Thomas’s overall customer experience is now dramatically different. If a billing error occurs, it is promptly identified, resolved, and a coffee voucher is sent to Thomas as an apology. A claim error will be flagged for proactive resolution, with the health plan provider working with the insurer to resolve the issue. Thomas no longer has to call the customer care line as often as he did before, and, because he has a hassle-free experience, his loyalty increases.
Now, when he gets sent a customer experience survey, he is happy to fill it out. A best-match offer is identified for him, and Thomas enrolls—through a personalized campaign—in a diabetes care program that improves his understanding of diabetes and adherence to care.
On the insurer’s side, the impact has been positive too. The next best experience capability has created lower cost-to-serve by minimizing inbound calls. It has increased revenue by lowering churn and improving cross-sell and upsell rates—all translating directly into bottom-line value for the company.
The next best experience capability uses data and AI to answer the question “What does this customer need most in this moment?” It then delivers a seamless, personalized, and satisfying customer experience that builds loyalty and customer lifetime value (CLTV).1 The AI-powered next best experience approach relies on data analytics, machine learning (ML)- and AI-powered predictive models, recommendation engines, and gen-AI content generation and personalization capabilities (see sidebar for an illustrative example, “How next best experience works: A journey to satisfaction and delight”).
Tackling customer engagement challenges with next best experience
As customer experience has become increasingly digital, managing customer relationships has become more complex. Often, multiple functions—billing, customer experience, customer care, and marketing—interact independently with customers, leading to poor results.
For example, the customer insights team might send a customer survey at the same time as the loyalty team pushes an email to sign up for automatic billing. Marketing, meanwhile, sends a new upselling offer. The customer ends up feeling spammed by the company, potentially ignoring all proactive communications or opting out of marketing consent altogether.
Next best experience can address these challenges by focusing on customer experience as a way to drive value, sequencing touchpoints, and using AI to drive personalization.
Refocusing on experience as the driver of value
The next best experience approach improves customer communication and engagement by providing a meaningful and personalized experience for existing customers. Deploying predictive analytics helps companies know when best to deliver the right messages.
Take, for example, a global payments processor that wanted to reduce attrition among its most valuable merchants: Using a next best experience approach, it built an advanced machine learning model to predict the likelihood of a merchant reducing business within the next seven days.
The model used a vast dataset, drawing on operational, financial, and customer information to build a digital twin of the daily interactions between the processor and each merchant. Once each merchant was scored, cluster analysis then grouped and prioritized the merchants by issue types (such as disputes) or opportunity (such as low working capital).
In the final step, the organization built a large library of interventions. These ranged from sales, such as introducing new products and features, to service, including fee forgiveness and technical fixes. It then mapped these actions to the clusters so that automated actions either protected revenue, reduced attrition, or capitalized on a client opportunity.
The result? The global payments processor estimated that the new system could reduce merchant attrition by up to 20 percent per year.
Sequencing touchpoints
By sequencing customer touchpoints more effectively, a next best experience approach can lead to more satisfied customers, potentially increasing revenue and reducing cost to serve.
For example, a telecommunications company in Europe decided to stop all outbound campaigns to customers who had open complaints, ongoing care journeys, or a high propensity to call regarding service-related topics. This simple act of ensuring that care activities took place prior to any outbound marketing had a positive effect: It drove the company’s net promoter score (NPS) to the same level as the market leader’s NPS and improved both cross-sell and churn rates.2
Deploying AI to drive personalization at scale
An AI-powered decisioning engine that governs decision-making across various use cases, such as new customer acquisition or personalized recommendations, can improve CLTV (Exhibit 1). Even better, an AI-driven next best experience approach improves with each new use case, as customer interactions are fed back into the integrated data set, making decisions more accurate over time.
A major US airline has harnessed AI for predictive customer insights to enable more personalized offers for high-value or at-risk customers. The customer service team did not previously differentiate between customers when offering compensation vouchers for flight delays or cancellations. However, by introducing machine learning models to inform recommendations, customer service agents could prioritize specific customer segments and tailor compensation accordingly.
In this way, the agents could differentiate, for example, between a frequent flyer who has faced three recent delays and a leisure traveler with no recent delays.
This AI-driven move led to a 210 percent improvement in targeting at-risk customers, an 800 percent increase in customer satisfaction, and a 59 percent reduction in churn intention among high-value, at-risk customers.
Constructing a next best experience engine
A next best experience engine can be built on an organization’s existing technology architecture, with an ecosystem of tech partners—including cloud, enabling technology, foundation models, and gen-AI solutions. The next best experience engine consists of four key components: data engineering, advanced analytics, gen AI, and a campaign delivery platform (Exhibit 2).
Data engineering
A solid data engineering foundation brings every customer record into a single repository, providing consistent insights. This process begins by pulling data into a data lake, gathering billing records, customer relationship management (CRM) entries, web analytics, mobile app events, and call center logs. Once ingested, these records are transformed into cleaned and aggregated tables using standardized scripts that prepare the data for analysis.
The tables are typically stored in modern data warehouses. Quality control processes are essential. Automated checks, for example, can detect anomalies (such as missing data or sudden drops in record volume). These controls are accompanied by data lineage tracking, which captures the origin, transformation, and update history of each data element, ensuring traceability for compliance and debugging purposes.
Advanced analytics
On top of the data foundation sits a series of models that predicts behaviors and recommends decisions based on the business context. These include three types of models: 1) propensity models that score how likely a customer is to upgrade, churn, or respond to a campaign; 2) channel models that determine whether email, text, in‑app messaging, or voice calls will be most effective; and 3) value models that calculate the lifetime value or near‑term revenue opportunity.
These models are interpreted through a decision orchestration layer that blends statistical outputs with real‑world operational logic. For example, a customer flagged as high-churn risk might be automatically removed from all promotional campaigns and instead added to a retention journey that involves loyalty offers or service improvements. Conversely, a customer with low-churn risk but high-upsell probability may receive a proactive upgrade message.
Gen AI—and agentic AI
Gen AI creates personalized content that powers the customer experience. Large language models enable dynamic message generation, which includes contextual references, such as recent transactions, service interactions, or usage trends. These messages can vary in tone and detail depending on customer preferences and can be tailored to the delivery channel (for example, SMS messages, personalized emails, or real-time chatbot conversations).
Initially, generated messages are typically reviewed by marketing or compliance teams. Over time, however, automated guardrails, template-based constraints, and learning loops will allow the content to scale with limited human oversight. More advanced implementations add agentic AI elements, where agents autonomously refine messages, predict impact, test different phrasing, gauge likely responses, and adjust accordingly. Agentic AI can move organizations beyond templated messaging to scalable, high-quality, personalized communication.
Campaign delivery platform
This final component ensures the delivery of messages through appropriate and timely customer channels, starting with integration into marketing automation systems that manage email sends, SMS notifications, and mobile push alerts. CRM systems and call center interfaces are also integrated, allowing human agents to view real-time prompts that reflect AI recommendations.
Using gen AI drives the ability to scale fast. It can be harnessed to create personalized content and be driven by recommendations from the central engine (for a company case study, see sidebar, “Using a next best experience engine to deliver an effective communication campaign,” and for an individualized case study, see sidebar, “Using AI to hyper-personalize customer communication”).
Despite the importance of getting the technology right, the successful implementation of a next best experience engine is never just about tech. It delivers value when it is embedded into workflows, supported by operational processes, and paired with organizational change. Even the most accurate model will fail if frontline teams do not trust or act on its recommendations.
This reality was clear in our own experience at McKinsey with the deployment of the internal gen-AI assistant, Lilli. More than 50 percent of the team focused not on technology but on change management, enablement, and training. Adoption was driven by embedding the tool across email, chat, learning platforms, and day-to-day rituals.3
How companies can get started on next best experience
Organizational leaders can consider six aspects when exploring the potential of a technology-enabled next best experience approach:
- The data layer: Organizations can start by integrating key data sources—such as CRM, billing, or operational datasets—into an initial feature store or sandbox data lake, and then check data accessibility and quality using a proof-of-concept ingestion pipeline.
- Advanced analytics: After selecting a high-impact use case, such as churn prediction or upsell likelihood, companies can train and deploy a basic predictive model in a pilot workflow. Performance can be monitored weekly to continually refine the model through feedback loops.
- A robust technology ecosystem: Companies can consider investing in technologies such as MLOps (machine learning operations), DevOps (development operations), and MarTech to streamline model development and deployment. This approach also helps better connect analytics and marketing for improved customer engagement.
- The operational model: Organizations can rethink their operating models, focusing on the importance of aligning incentives across silos. For example, a single contact policy across all teams can drive coordinated customer interactions across different departments, such as billing, customer experience, marketing, and sales. Implementation can include establishing a cross-functional working group with a unified contact policy and aligned incentives, such as shared targets, to drive collaboration.
- Impact measurement: Companies can set up a universal control group and a universal target group. Comparing the two can reveal insights into customer behavior and transaction patterns. At the same time, control groups can be used for A/B testing of campaigns, while dashboards provide transparency into predictive model performance.
- A two-speed approach: Companies can consider launching a small-scale lighthouse pilot while concurrently working on foundational capabilities, such as data lake architecture and cross-functional governance. This approach can drive rapid results alongside a scalable build-out.
The convergence of technological progress and heightened customer expectations has created an ideal environment for revolutionizing customer engagement. The next frontier of the customer experience—the next best experience approach—can now be unlocked, especially by front-runners who can effectively use AI and gen AI to enhance predictive capabilities and gain a competitive edge.
