Case Study: How we redesigned a fintech onboarding process using AI, without losing the human touch

When FinanceFlow (name changed for confidentiality) approached us, their digital onboarding was hemorrhaging potential customers. Despite having a solid product, they were losing 40% of users during sign-up. The culprit? A clunky, impersonal process that took weeks to complete.

Their CEO was clear: "We need speed, but we can't afford to lose the trust factor that sets us apart from big banks."

The Challenge: Speed vs. Human Connection

FinanceFlow faced the classic fintech dilemma. Their manual onboarding process involved multiple touchpoints, document reviews, and compliance checks that could take 14-21 days. Meanwhile, competitors were promising instant approvals.

But here's the thing, their customer research showed that trust trumped speed. Users wanted efficiency, yes, but they also wanted to feel seen and supported, especially when handing over sensitive financial information.

Our Hybrid Approach: AI as Assistant, Not Replacement

Instead of throwing AI at every problem, we took a more strategic approach. We mapped the entire journey and identified where automation could eliminate friction without eliminating humanity.

What we automated:

• Document verification and data extraction

• Initial compliance screening

• Risk assessment calculations

• Progress tracking and status updates

What stayed human:

• Complex application reviews

• Customer questions and concerns

• Edge cases and exceptions

• Relationship building conversations

Live Co-Creation Workshops: Remote + Hybrid

I ran fast, focused sessions.

Who joined:

• Product

• Compliance

• Customer Support

• End Users

Energy:

• Mics on. Hands up. Cameras steady.

• Sticky notes everywhere. Chat on fire.

Human first:

• Empathy maps. Said aloud.

• Inclusive brainstorming. Silent start. Round-robin.

• Accessibility checks. Every sketch. Every flow.

AI in the room:

• Instant insight from past onboarding data.

• Drop-off clusters auto-surfaced.

• Idea prompts. Variations. Mashups.

• Real-time copy nudges. Plain language.

Moments:

• “We decided in five minutes.” — Compliance Lead

• “That screen can go.” — Support Agent

Rhythm:

• Hybrid boards mirrored. Roles mixed in breakouts.

• Tight timers. Clear votes. Done.

The Solution: Smart Automation with Human Safety Nets

Based on our research, we redesigned the onboarding flow with three core principles:

1. Progressive Disclosure

Instead of overwhelming users with forms, we broke the process into digestible chunks. AI algorithms determined the optimal order based on user behavior patterns, while keeping humans available for questions at each stage.

2. Proactive Support

We implemented AI chatbots for instant responses to common questions, but with seamless escalation to human agents. If someone hesitated for more than 2 minutes on any step, they'd get a gentle nudge offering help.

3. Transparent Progress Tracking

Users could see exactly where they were in the process, what came next, and realistic timelines. This reduced anxiety and support tickets by 35%.

Implementation: The Technical and Human Elements

In-Workshop AI

• Instant insight from past onboarding data by segment

• Idea generation with smart prompts and mashups

• Rapid prioritization via live voting + impact/effort scoring

• Real-time decisions on flows and wording with previews

• Plain-language rewrite suggestions for microcopy

• Compliance risk cues on sensitive terms

Human Moves

• Empathy mapping across roles before sketching

• Inclusive techniques: 1-2-4-All, round-robins, silent starts

• Accessibility checks: screen reader, contrast, keyboard-only

• Microcopy read-alouds to catch tone

• Edge-case walk-throughs

• Decision log captured on one board

The Results: Numbers That Tell a Story

Before / After. Clear.

“Built with me, not at me.” “I knew what was next. I stayed.”

Lessons Learned: What Worked (And What Didn't)

What Worked Brilliantly:

• AI for speed, humans for complexity proved to be the sweet spot

• Proactive communication reduced anxiety and abandonment

• Contextual help based on user behavior was more effective than generic FAQs

What We Had to Adjust:

• Initial AI responses were too formal: we had to train them to be more conversational

• Some users preferred phone calls over chat: we added callback options

• Complex financial situations still needed human expertise from day one

Workshop Takeaways for Fintech Service Design Teams

• Co-create live. Remote + hybrid. Mix roles.

• Use AI for insight, ideas, priorities. Keep humans for nuance.

• Decide in the room. Log decisions. Ship small. Iterate weekly.

• Lead with empathy maps. Then flows, then copy.

• Bake in accessibility. Test every screen.

• Keep the human escape hatch. Always.

Summary: Build together. Move fast. Stay human. Stay Tuned.

The Bottom Line

FinanceFlow's transformation proves that you don't have to sacrifice humanity for efficiency. By using AI to eliminate busywork and friction, we freed up their team to focus on what really matters: building trust and solving customer problems.

The key insight? Technology should amplify human capabilities, not replace them. When you get that balance right, everyone wins: your customers get better service, your team gets more fulfilling work, and your business gets better results.

Ready to redesign your fintech onboarding process? At Blue Tango Design, we specialize in blending cutting-edge AI with human-centered design. Let's build something that your customers will actually love using.