AI should make banking faster and simpler without losing the human connection that builds trust. In our recent American Banker webinar, Tom Martin (CEO, Glance) and Josh Johnson (Sr. Director, Operations Strategy, SoFi) shared how leaders can blend AI efficiency with human empathy to earn loyalty, increase adoption, and deliver better outcomes.
Here are the key takeaways.
Why “humanizing AI” matters
Banking customers expect two things at once: speed and empathy. AI is great at removing friction: surfacing answers, automating steps, and handling repetitive tasks. But when conversations get complex or emotional (fraud, disputes, major financial decisions), humans build the trust that keeps customers loyal. The goal isn’t AI or humans; it’s the right orchestration of both.
“I think banks are recognizing that they need both (AI and humans). They can't go with one versus the other. And it's that balance of what are the types of tasks that really we want to create efficiency, what humans expect, but the ones that really lead to needing more empathy, let's make sure we connect them with a human being without hiding them or making it too difficult.” - Tom Martin, CEO, Glance
5 insights leaders can use now
1) Draw the line at trust
Efficiency shouldn’t come at the expense of confidence. Use AI to streamline routine tasks and route high-stakes moments to people, fast. Hiding the human or making escalation hard erodes trust.
2) Govern AI like a critical system
Before production, validate models and experiences, define escalation triggers (e.g., fraud flags, security issues), and maintain strong feedback loops. Governance prevents “endless loops” and protects customers and the brand.
3) Make real-time help feel effortless
Give customers control over channel and pace. When a handoff to a person is needed, make it instant and context-aware so no one repeats themselves. In-session tools like cobrowse, screen share, secure doc exchange, and video help resolve complex tasks quickly and clearly.
4) Rethink the agent experience
AI should free agents from drudgery (summaries, notes, lookups, compliance prompts) and equip them with context and next-best actions. As AI handles simple contacts, agents will spend more time on complex, empathy-heavy work, so invest in training, change management, and the right tooling.
5) Measure what matters (and expect shifts)
Expect AI to reduce volume and speed up simple interactions. Human-handled cases may take longer, but should deliver higher first contact resolution (FCR), better CSAT/NPS, and stronger customer effort scores. Track loyalty and product adoption too; the goal is lifetime value, not just faster calls.
Spotlight: SoFi’s approach
As one of the most recognized names in digital finance, SoFi has built its reputation on making banking simple, fast, and human-centered. During the webinar, Josh Johnson explained how SoFi introduces AI with transparency and care:
- Customers know when they’re interacting with a bot.
- Escalation to a human is seamless when needed.
- AI is applied only in carefully selected use cases - those that reduce repetitive tasks while preserving trust.
“We don’t want to make it transactional. We’re transparent when members are talking to a bot, and if they need to get to an agent, it’s a very seamless experience. We are focused on using AI to automate those really repetitive tasks that are prone to human error or take a lot of time to do. And we are really very selective on the use cases where we're applying AI to ensure that members feel that the safety and trust is there from a banking institution. When you're dealing with people's money, you want to be very sensitive to that and make sure that your members feel very safe in their interaction with the group.” - Josh Johnson, Senior Director of Operations, Strategy, SoFi
This strategy ensures SoFi can scale efficiently while maintaining the safety, empathy, and reliability its members expect.
Read the full SoFi + Glance case study here.
The near future: seamless, predictive, human-centered
Customers will expect faster, more accurate, more self-serve experiences. When they do reach out, they’ll want one-touch access to a person who has the full context and the tools to help right now. Think of AI as a concierge: instant on the front end, and empowering for agents when a handoff happens.
Practical playbook
- Start small, learn fast: Pilot targeted use cases, instrument feedback, and iterate.
- Make escalation obvious: Be transparent when it’s a bot and offer a clear, one-tap path to a human.
- Invest in agent enablement: Cobrowse, guided workflows, knowledge surfacing, summaries, and secure file exchange.
- Design for low effort: Fewer steps, zero repetition, and context carried across channels.
How Glance helps
Glance embeds human-guided collaboration, including cobrowse, video, screen share, and secure document exchange, directly in your website/app and CRM. That means agents can instantly see what the member sees, guide complex tasks, verify documents, and connect face-to-face when it matters most. No downloads, fully brandable, enterprise-grade security, and proven at scale.
.png)