Like many businesses, we're exploring how to integrate AI into our technology products. But here's the difference—we're not overpromising or trying to solve every problem all at once. Instead, we've embraced a pragmatic, actionable approach that’s producing real results today—not in some distant, hypothetical future.
This blog series will take you behind the scenes of how we’re leveraging AI in three critical areas:
- Accelerating time-to-insight in data analysis and portfolio management
- Enhancing recommendation engines by combining quantitative models and qualitative judgment
- Speeding up internal statistical modeling in ways we plan to extend to our client
Why Most AI Projects Fail
Here’s an uncomfortable truth: 85% of AI projects fail because their ambitions outstrip their capabilities. Teams oversell what AI can do, only to deliver solutions that fall short.
I’ve seen this happen time and again. Take the forklift company that claimed AI handled 95% of their forklift operations. Impressive, right? Except they still needed humans for the remaining 5%, tasks too complex for their AI to manage. Their aim was full automation, but they failed to get there. Or consider the coding company that claimed they had AI-generated code but really employed 7,000 engineers outsourcing work under the pretense of artificial intelligence.
The pattern is predictable: companies envision overly ambitious responsibilities to AI, hoping for revolutionary outcomes but fail due to over-promising and under-delivering.
A Reality Check on AI
AI isn’t a miracle worker. It doesn’t create data out of thin air or uncover truths that are completely unknowable. What AI does excel at is processing vast amounts of data faster than humans can. That’s its real “superpower.”
The claims that AI “finds insights no human could imagine” are exaggerated. The truth is, AI builds and tests models faster, so it identifies more complex relationships and patterns faster. A human could theoretically reach the same insights, but it might take them years instead of hours.
This distinction matters, especially in financial services, where tight compliance environments demand decisions that are explainable and rooted in human understanding. Decisions made by systems that no one fully understands aren’t just risky; they’re unacceptable.
Our Crawl-Walk-Run Philosophy
We know where we want to go with AI at Nomis. The ultimate vision? Seamlessly interacting with our tools using natural language, asking questions and getting answers effortlessly. It’s an ambitious goal, but not one we’re chasing without strategy or caution.
Why? The risks are real. The development costs are high, the technical challenges are immense, and issues like hallucinations and implementation pitfalls can’t be ignored.
That’s why we’re starting with clear, focused use cases that create value right now. Here’s how we’re using AI today:
1. Surfacing Existing Insights Faster
Our products already deliver comprehensive data analysis for portfolios, product structures, and marketing campaigns. These tools empower product managers with insights that were previously out of reach.
Now, we’re using AI to accelerate this process. What used to take weeks now only takes minutes, dramatically reducing the time to valuable insights.
2. Supporting Better Decisions with Natural Language
Commercial relationship managers, bankers, and branch managers spend significant time gathering information before making decisions. Our AI does that groundwork for them.
Using customer data, statistical models, scorecards, and business policies, we provide recommendations in seconds. This isn’t about replacing their expertise; it’s about making their decision-making faster and easier.
3. Accelerating Model Evaluation
At Nomis, we pride ourselves on custom-built price elasticity models tailored for each client. These aren’t generic "industry-standard" models. Each one is unique to the customer data at hand, which means we need to create and test a large number of models.
Here’s where AI helps. It accelerates our model evaluation process, pinpointing areas where human expertise should focus. Crucially, we don’t rely on AI to verify its own work. Humans always have the final say, ensuring quality and trustworthiness in everything we deliver.
Our Non-Negotiable Principles
Two guiding principles shape our approach to AI.
Data Security Comes First
Protecting client data is our top priority. Every decision we make starts with ensuring that sensitive information remains safe.
Humans Stay in the Loop
We’re not automating decision-making. Instead, we’re using AI to empower the people who drive businesses, giving them better tools to make smarter, faster decisions.
What's Next
Over the coming weeks, we'll share detailed deep dives into each of these use cases. You’ll see real examples from our tools, learn how we designed these solutions, and understand why we believe our approach sets us apart.
Our goal isn’t to replace human judgment with AI. It’s to augment human intelligence with the acceleration that AI provides. That might not sound as flashy as “self-governing AI,” but unlike many overhyped promises, our approach works. And at the end of the day, isn’t that what really matters?
Stay tuned for Part 2: How AI Accelerates Data Analysis Without Replacing Analysts
If you’re rethinking your approach to AI utilization, Nomis can help. Get in touch with the team at sales@nomissolutions.com or connect with us via nomissolutions.com to speak directly with our experts!
Written by: Wes West, Chief Analytics Officer at Nomis Solutions