For AI in Fintech, User Experience Still Matters

Female and male warehouse employees standing in the aisle looking at a tablet computer


By Shaun Illingworth

“What do you do?” can be a tough question to answer. I have trouble with it sometimes, and I’ve been essentially doing the same job for decades. It can also be a tough question for a start-up tech company. The core personnel are so immersed in, and excited by, the capabilities of the product it can be difficult to take a step back and elucidate how it will help potential clients.

A woman works on laptop computer while flipping through a stack of documentsFor Vigilant AI, an Ottawa-based start-up in the financial services sector, clarifying how clients would use their breakthrough data analysis technology was one challenge in the early stages of development. Another challenge was communicating their value proposition to potential investors.

DFFRNT came on board to assist with both these tasks. The principals at DFFRNT have extensive experience with technology for the financial services sector and we were excited to apply human-centred design principles to a product that incorporates artificial intelligence tools.

Artificial intelligence is spreading tendrils into every aspect of business and government. It is most useful, of course, for tasks that require analysis of large amounts of data. It would be a perfect fit for financial auditing functions, except that financial data is often not in appropriate structured formats.

Vigilant AI is tackling that challenge. Vigilant AI has pioneered a platform with an intuitive data collection process which allows an AI-fueled tool to ingest and analyze documentation and financial data involved in the supply chain more quickly. This supports quicker detection of financial loss and provides unprecedented supply chain insights.


Our role was to translate complex business intelligence and data analytics capabilities into an engaging user experience and make the product’s value clear to investors

John Craig, Vigilant AI’s CEO, believes the unique approach of combining both structured and unstructured financial data into a fully contextualized data lake will help auditors reduce risk, increase margins and create new client service offerings.

Vigilant AI asked us to visualize the user experience for their new suite of products and provide an early prototype. We started with conversations with the team and scribbled notes on a on a white board. We discussed what the service was able to do and the results it could provide. Our team applied its considerable research and design skills to the project, and a few weeks later, we provided Vigilant AI an interactive prototype.

We also helped this tech startup articulate its value proposition. It was important to capture the interest of an investment audience, so our role was to translate complex business intelligence and data analytics capabilities into concepts an investor would appreciate.

AI offers completeness, speed and cost savings

In the course of our work with Vigilant AI, John and I have had many conversations about the suitability of AI for financial tasks and the imperative to keep user needs front and center with AI-powered products.

John is convinced AI will be the driver of the next industrial revolution.

For this blog, I sent him a few specific questions to get both sides of the product development story.

Asked why Vigilant AI chose DFFRNT, John said: “DFFRNT speaks the language of user-centric design, which neatly filled a hole in our early-stage organization, allowing us to bring our team under a singular vision of how we wanted the product to interact and display information to our target audience. This allowed the development team to focus on the “how” they were going to build the architecture to deliver on the vision, and enabled our founding team to capture the attention of investors who could see how our platform would accelerate how they analyzed structured and unstructured data at the same time.”

We were able to step back from the complicated math and analytics that underly Vigilant’s product suite, and help the Vigilant team focus on the benefits to the client. We translated the value of these sophisticated, complex datasets into easy-to-understand, actionable terms and features that appeal to potential users.

John says there are three key benefits to using AI in financial analysis:

  • completeness,
  • speed, and
  • cost of operations.

“The first benefit is completeness. Just to provide context, PWC estimated that in 2018, 5 years ago, 18 Zetabytes of financial information was created globally, of which 10% is structured financial information, and 90% is unstructured documents, such as emails and legal agreements. With numbers of that size, current accounting methodologies, such as random sampling, no longer provide reliable review capabilities. AI provides the only way to support auditors in reviewing 100% of the available data during tight audit timelines.

“The second benefit is speed. AI works faster than humans, but to work optimally, you need a well-contextualized data lake of information to review, and that’s what we’ve focused on delivering in our first phase of development, positioning us to work with the latest and greatest in AI approaches as they become available, including ChatGPT.

“The third benefit is cost of operations. We reduce the labour required to review data, helping to increase the auditor’s margins.”

Searching for value in unstructured data

The unstructured data John refers to are notes and information not contained within a database framework. They may be contained in the text of an email or a PDF document. Previously, auditors would have to manually search these sources. For example, if a buyer receives a special deal on shampoo for a limited time, this information may only be stated in an email and not captured in purchase orders and accounting systems. If this information is not conveyed to the accounting systems, there is a financial loss associated with the missed opportunity for the buyer’s company.

The data intelligence tool within Vigilant AI’s platform is able to search unstructured data for this type of occurrence, but there’s still the challenge of gathering the source data.

John explains: “One of the drawbacks of our approach is data acquisition. Companies under audit rarely have all the required data for an audit in one place. The legal department owns master service agreements with suppliers. IT manages the email system. The office of finance generally has their structured financial information in a cloud-based ERP. We have tackled this issue by enabling information holders in each department to publish their own information to the data lake, speeding data collection.”

Working with Vigilant has made us more aware of how AI, machine learning and large language models will intersect with financial services. The large volumes of data involved with financial technology make it almost a necessity.

There is absolutely a place for AI in financial services technologies, but if these tools are to reduce the workload on humans, developers must keep user needs front and center.

Related blog posts

In the era of online banking, clients still value the personal relationship with their local branch. They appreciate knowledgeable staff, ask tellers and advisors for input on financial decisions, and overwhelmingly choose face-to-face meetings or phone calls when they have a question or concern.
What started as brown bag “lunch and learn” sessions over 15 years ago have evolved into a full program of corporate and academic teaching – and we love it!
We are visionary, intelligent, funny and not afraid to speak out because we are backed by years of experience and knowledge. We are DFFRNT.