By 2026, things were getting dicey for Ascent Financial. Their costs to acquire a customer were going up, and the retention rates they used to brag about were flatlining. Staring at the Q1 report, Ascent’s VP of Marketing, Sarah Chen, felt that familiar pit in her stomach. “We’re just throwing money at siloed campaigns,” she told her team, “and every time a customer interacts with us it’s like they’re starting from scratch. How do we build a single, coherent customer experience that feels like we’re one step ahead of them?” The challenge was clear: they had to design proactive agentic customer journeys, but figuring out how to do it felt like hacking through a dense jungle with no map.

For a regional bank in the Pacific Northwest, Ascent Financial had already sunk a lot of cash into digital transformation over the last three years. They had a decent mobile app, a clean online banking portal, and even a chatbot on the website. The problem was, none of these tools talked to each other. It was the classic story: a customer starts a loan application online, calls customer service with a question, and has to explain the whole thing from scratch to an agent who has zero context. Then a marketing email would hit their inbox promoting a savings account, completely oblivious to the mortgage application they were in the middle of. This kind of disjointed experience was annoying customers, and worse, it was costing Ascent real money.

Sarah knew the problem wasn’t the tech itself, it was the missing orchestration. “A customer is talking to *Ascent Financial*,” she’d say in strategy meetings, “not a website or an app. The whole journey has to feel like one intelligent conversation, not a bunch of disconnected pings.” This meant a jump to genuine agentic design, where the system actually anticipates what a customer needs and guides them, instead of just reacting to clicks. This fundamentally changes how you think about engagement.

Job one was getting all their customer data in one place. Information was all over the map, stuck in their CRM (Salesforce Service Cloud), their marketing automation tool (HubSpot), and the company’s core banking system. As Sarah put it, “We’re building on sand without a single customer view.” They decided to bring in a customer data platform (CDP), a massive six-month project that became the company’s central nervous system, pulling in data from every single interaction to build a real-time profile for each customer, a task that involved a dedicated cross-functional team and some intense, painful work on data hygiene and mapping. It was a beast.

With the CDP humming, they could finally start designing the actual agentic journeys. Sarah’s team went after the highest-friction areas first, like the onboarding for a new checking account, which at the time was just a clunky series of emails and prompts that felt more like a chore than a welcome.

“We have to think in states and transitions,” David, Ascent’s lead UX designer, kept saying. “What does someone *actually need* right after opening an account? They don’t need a blast of information, they need guidance. What’s their next logical move, and how do we give them a gentle nudge in that direction?” They mapped out the ideal path, spotting the key decision points. For instance, after opening an account, most people need to set up direct deposit or link another bank. So instead of a generic “check out our features!” email, the new agentic journey would see the account was opened, send a quick personalized welcome, and then proactively offer to walk them through setting up direct deposit, a guided flow right in the mobile app that even pre-filled information where it could.

Their conversational AI got a major upgrade, too. The old chatbot was just a glorified FAQ bot, totally useless if a conversation spanned more than one session. The new AI was trained on piles of customer interaction data and, more importantly, plugged directly into the CDP. This meant if you started a loan application online, bailed, and came back to the chat later, the bot wouldn’t just ask “How can I help you today?” It would say something like, “I see you were working on a personal loan application. Can I help you complete it or answer any questions?” This small change made interactions feel personal, not transactional. It tracks with what others are seeing. A Nielsen report from early 2024 showed that companies with predictive support get a 15% bump in customer satisfaction scores over those who are purely reactive.

They also wired up real-time behavioral analytics. So if a customer was stuck on the ‘forgot password’ page for more than two minutes, for example, the system didn’t just sit there. It would automatically fire off a personalized email with a direct reset link or even pop a live chat offer. This was about providing timely, relevant assistance right when it was needed. As Sarah noted, the goal was “anticipating friction and smoothing it out before it becomes a full-blown problem.”

Getting consistency across channels was tough. An agentic journey falls apart if the experience on the mobile app is totally different from a phone call. Ascent had to do a ton of training with their customer service representatives, who now had access to the same unified customer view as the AI. They were no longer just putting out fires. When a customer called, the rep could see their entire recent history, abandoned forms, chatbot conversations, everything, which eliminated that infuriating “let me get you up to speed” dance. “We had to retrain our reps to be orchestrators,” Sarah explained, “to see how their one conversation fits into the customer’s larger story with us.”

Within a year of the full rollout, the numbers started talking. Average customer service call times for common inquiries dropped by 20%, mostly because the AI could handle the simple stuff and human agents had all the data they needed. Their Net Promoter Score jumped 8 points. New account activations, particularly for their digital-first products, also climbed by 12%. The investment in the CDP and advanced AI was substantial, but the return on efficiency and customer loyalty made the case.

This kind of project is never really “done.” It requires continuous effort to understand and anticipate what customers actually need. You absolutely need the unified data, smart AI, and a company culture that’s obsessed with a proactive, smooth customer experience. For Ascent Financial, this whole effort changed their customer interactions from a series of disconnected events into a single, intelligent conversation, which is what actually ends up strengthening relationships with clients.