Case Study: Bank of America’s ROI on AI

-


 



How does AI generate value?


Today’s post springs from an episode of Bloomberg Technology that put a spotlight on Bank of America’s AI-powered virtual assistant, Erica.

Billions are being poured into artificial intelligence. But here’s the hard truth: an MIT study found that 95% of corporate AI projects fail to deliver measurable financial returns. 

Only 5% make it past the pilot stage with a real boost to the bottom line. 

So when Bank of America’s Erica racked up billions of interactions and over $1 billion in value, it’s worth paying attention. What did they do differently—and what can Caribbean banks learn from it?


Erica by the Numbers

Since 2018, Erica has:

  • Surpassed 3 billion client interactions with tens of millions of customers.

  • Pushed millions of proactive money alerts—from subscription hikes to unusual spending.

  • Cut employee IT help-desk calls in half, with >90% staff adoption.


Why Erica Works When Others Fail

  • Real pain points first: Customers want quick answers; employees want faster IT support.

  • Automation + human touch: Routes to live agents when needed, avoiding the dreaded “chatbot wall.”

  • Scale carefully: No risky open-ended AI for customers—trust and accuracy come first.

Think about it: a bank isn’t just about numbers—it’s about trust. Your money shouldn’t vanish into a digital black hole, and Erica makes sure customers see what’s happening and why.


What Caribbean Banks Can Do Differently

Caribbean banks don’t have the luxury of billion-dollar budgets—but small markets are perfect for focused, high-impact pilots. Here’s how to start:

1️⃣ Build Trust Before Experimenting

Start with AI that gives deterministic answers—like a GPS, it always takes the same predictable route. Only after that, try open-ended chat.

2️⃣ Hyper-Personalized Banking & Education

AI can predict seasonal cash spikes—Carnival costumes, Christmas shopping etc.—and suggest custom savings plans or micro-loans. It can also give bite-sized tips like:

"You’re at 80% of your US-dollar limit—time to ease up!" (joking- go ahead and buy that plane ticket)

3️⃣ Customer Guidance & Alerts

AI can notify customers about required documents, expected timelines, or potential delays. This makes the process more transparent, so users know exactly what’s happening.


The Bottom Line

For Caribbean banks, the opportunity is not to copy-paste BofA’s model, but to start with pain points that matter most in the region—like credit card usage, FX monitoring, and compliance—then scale up with trust. Because at the end of the day, a bank’s most valuable asset isn’t AI—it’s the confidence customers have that their money is safe.


Resources

Comments

Post a Comment

Popular posts from this blog

0 to 100: A Reflection

-
Image
What's 10 x 10? I started this blog on February 4th, 2023—almost two years ago (wow, I just checked!)—and I must say, writing consistently is a challenge that I gladly took on. Writing has been a constant theme in my life. It got me through my rebellious teen years and straight into adulthood (which, to be honest, I still have a hard time accepting). Many people ask me, "How do you do it?" To that, I say: I enjoy it, and I have a great support system. Today, I want to take a moment to reflect and give thanks—to those who read my posts, those who point out my typos, and especially to my dad, who always asks, "Toni, did you post yet?" But let’s be real: doing 100 of anything is difficult—whether it's 100 pull-ups or developing 100 apps. Even when you’re passionate about something, there are days when you just don’t have the energy or inspiration. And that’s okay! You just have to keep pushing forward—through the valley of despair, even when no one is watchin...
Read more

Learning Something New: EDA on Guitars

-
Image
This is an acoustic guitar Today’s post is inspired by my childhood dream of learning to play the guitar. For years, I’d daydream about strumming chords and creating melodies, but it always felt like something I’d do “someday.”  That day finally arrived when I jokingly shared this dream aloud, and my dad decided it was time to make it a reality. Yesterday, he surprised me with my first guitar—a sleek, acoustic guitar with nylon strings. The moment I picked it up, I was both excited and intimidated. I immediately started tuning it and spent an hour practicing E-minor and A-minor chords.  My fingers quickly grew sore, and faint ridges were left behind from pressing on the strings—but I didn’t mind.  There was something unexpectedly soothing about the vibration of the strings resonating through the guitar, as if the instrument itself was encouraging me to keep going. Learning a new skill is a reminder of how persistence pays off. I know that mastering the guitar will take ti...
Read more

Make Your Screen Time Matter

-
Image
People swipe on their phones an average of 2500-5000 times per day   With the TikTok ban looming in the U.S. amidst national security concerns, I believe we should also be asking a deeper question:  What are we allowing to capture our attention? Social media isn’t just profiting from our data—it’s also draining our limited time and energy! Isn’t it funny how we all wonder why we can’t get things done, yet we constantly allow ourselves to be distracted? As someone who recently deactivated their Instagram account, I’ve realized how much passively consuming endless videos— ones I never even searched for —has negatively impacted my well-being. Whether it was comparing my life to those who could afford trips to Europe or staying up late searching for something to share with friends, I realized I was consuming content without purpose. In a time of shrinking attention spans, the ability to focus and take action is more valuable than ever. Now, I’m not saying entertainment is bad, but...
Read more

Scheduling Algorithms

-
Image
Hickory Dickory Dock! Today's post is inspired by my sister (AKA Dr. Dan Dan), who recently asked me for help scheduling appointments for her patients. As a budding dentist, it's imperative for her to find patients and fit them into her already busy schedule. Sometimes, patients cancel, and she faces a swiftly approaching deadline. Just like a computer prioritizes tasks based on system resources, my sister must prioritize her patients based on the complexity of the procedure and the time available. For example, root canals require more time and expertise than a simple cleaning, much like certain programs demand more of a computer's memory and processing power. The Task Manager in your computer is a program that manages all the active processes, or tasks, running at any given time.  It prioritizes tasks based on importance, ensuring that the most immediate requests—like opening Google Chrome—are allocated more system resources. With that being said here are some common sched...
Read more

Sharks, Dogs and Biases

-
Image
Tens of millions of people suffer from dog bites each year globally, compared to just 50–80 shark bites. As a data professional, I am expected to interpret data and provide stakeholders with insights in the form of a story. Many refer to this skill as "critical thinking." While that is valuable, one must also question how they arrived at their conclusions in the first place. According to Wikipedia, a bias is a disproportionate weight in favor of or against an idea or thing, usually in a way that is inaccurate, closed-minded, prejudicial, or unfair. I gained a deeper understanding of biases during an Organizational Behavior course I took in university, and I briefly mentioned it in Blog Post #8, which discusses data-driven decision-making. Beyond numbers in a database, biases can lead to poor decisions in the real world. For example, it is statistically more likely to be bitten by a dog than by a shark; however, availability bias caused authorities in Tobago to place a bounty ...
Read more

Value Creation

-
Image
Money grows on trees—the tree is a metaphor for your mind. Defining a problem is the first step to solving it, but have you ever considered the value your solution will create? Before jumping into problem-solving, it's crucial to ask: Is this the right problem to tackle, and is solving it worth the cost? As professionals—especially in tech—it’s easy to fall into the habit of building solutions without pausing to assess whether the effort aligns with the value it delivers. For example, imagine spending weeks building an automated self-watering system for a cactus. Sure, it’s a cool project, but was it really worth the time and resources? After all, a cactus only needs watering once a week, and it would take less than a minute to do it yourself. Sometimes, solving the problem manually is the most effective and resource-efficient option. Your salary is directly linked to the value you bring to a company, and time and energy are finite resources. How we choose to allocate them signif...
Read more

The Algorithm : Musk's Mental Framework

-
Image
A Decision Tree Today’s post is inspired by Elon Musk, the captivating biography by Walter Isaacson. At 688 pages, it’s a deep dive into the mind and methods of one of the most transformative figures of our time. Over the past two weeks, I’ve spent my early mornings absorbing Musk’s journey—and it’s left me contemplating the driving force behind his success.  What is the “algorithm” that powers his extraordinary ventures? Though his leadership style is unconventional, it’s undeniably effective. Musk has co-founded revolutionary companies like PayPal, SpaceX, and Tesla, each of which has disrupted its industry and pushed boundaries.  But what makes him tick? What’s the underlying formula behind his ability to juggle ambitious goals, relentless innovation, and transformative leadership? The Algorithm: Musk's Mental Framework As detailed in Isaacson’s biography, Musk often speaks of a personal algorithm—a mental framework that guides his decision-making process.  The algorit...
Read more

Understanding Outliers

-
Image
You can clearly see which one stands out   Today's post is inspired by a question one of my colleagues had around outliers: Are there different types? And to that question, the simple answer is yes! As data professionals, we often encounter data points that stand out like a sore thumb during analysis, leaving us wondering what to do with them. In most cases, deletion seems like the obvious choice.  However, I believe we need to pause and take a closer look at these points before removing them. You see, during Exploratory Data Analysis (EDA for short), asking the right questions about the dataset is crucial to uncovering its full story.  Deleting an outlier just because it stands out can create a narrow, myopic view of the data, limiting the depth and value of the insights you can gain. Rather than rushing to remove outliers, consider what they might reveal. Could they signal an emerging trend, an unusual but important event, or even a data collection issue?  With t...
Read more

Data-Driven Case Study : Barnes & Noble

-
Image
Barnes & Noble's beginnings can be traced to 1873 Today's post is inspired by an article I read about the recent resurgence of Barnes & Noble, the iconic American bookstore chain. As a bookworm, I found the story especially appealing. Back in secondary school, I used to imagine working in a bookstore or a library during summer vacation, surrounded by shelves of book. It seemed like a dream job. Barnes & Noble originally set itself apart with vast, well-stocked stores and an enormous selection of titles. This approach attracted a steady stream of customers and fueled the company’s rapid expansion. But then Amazon arrived, making it possible to order a book with a few clicks and have it delivered to your doorstep. The shift to online shopping left many of Barnes & Noble's stores eerily empty. Enter James Daunt, the managing director of Waterstones, the largest bookseller in the United Kingdom. Tasked with turning Barnes & Noble around, he brought a fresh p...
Read more

Data Stacks: Google, Microsoft and Amazon

-
Image
Stack: a pile of objects, typically one that is neatly arranged. Conversations about tech preferences tend to get heated, whether it’s Apple vs Android, Windows vs macOS, or the eternal debate about pineapple on pizza. But lately, I’ve found myself drawn to something : data stacks. Specifically, why do organizations choose Microsoft, Amazon, or Google as their cloud platform? Microsoft has quietly become the stack of choice for many enterprises, but why? What’s the real difference between AWS and Azure, and is one easier to learn than the other? More importantly, are we so locked into these proprietary tools that we’re missing out on better, open-source alternatives? Learning a new tool takes time, and that time investment often keeps us tethered to one ecosystem. But with so much innovation happening, it’s worth asking: Are we keeping our minds open to what’s out there? To answer these questions, it helps to ...
Read more