Artificial intelligence is transforming the way businesses operate, and many are seizing the opportunity to gain a competitive edge through automation.
Although forward-thinking businesses are embracing AI to streamline their operations, others are approaching this emerging technology with caution. And by doing so, they may be overlooking the potential benefits, especially if they continue relying on outdated systems and manual processes.
In a recent PaymentsJournal webinar, Ahsan Shah, Senior Vice President, Data Analytics, at Billtrust, and Christopher Miller, Lead Analyst of Emerging Payments at Javelin Strategy & Research, delved into just how far AI has come over the past few years, particularly in the realm of generative AI and deep learning, and how businesses can successfully leverage AI within their operations.
AI’s Evolution
Shah, in describing AI, likened it to an onion. AI is the outermost layer, the broader ecosystem. Other key components—including machine learning, deep learning, natural language processing, and generative AI—can be found below, under deeper layers. And in the past five to six years, there’s been more of a focus on deep learning and generative AI.
“Everyone’s talking about how generative AI will help, and that is where you essentially have language models, foundational models built by large companies like OpenAI, Google, and Anthropic,” Shah said.
“What this has done, which is a bit different than the other layers of the onion, is give you a language-based interface, a multimodal interface to say I speak the language, and then it can translate that. I can even feed it an image—it can recognize the image and allow you to generate more personalized content. It’s almost like a library of Alexandria. You don’t need to give it your own data, but now you have this interface within the world of AI that gives you another toolkit to do very amazing things.”
AI is just one component when it comes to developing customer and product value from data. Other components include traditional transactional reporting and analytics, all of which create multifaceted layers of value.
Although AI is a powerful technology, it shouldn’t be taken as the be-all, end-all solution. Analytics remain an indispensable component that organizations rely on to make more informed decisions. Thus, OpenAI’s ChatGPT should not be utilized in isolation.
Two key takeaways are emerging, Miller says. The first is the imperative for a shared data ecosystem that facilitates seamless implementation. The second is the potential of generative AI to automate tasks.
“Generative AI creates a move up the value chain in terms of what types of decisions or functions can be automated,” Miller said. “So where report creation might have been a very manual process, we can start to automate the creation.
“The information can be updated in real time as opposed to once a week when someone has to download an Excel file, run a series of macros, and add some data to a slide that gets sent to somebody else and presented in a report.”
Shah also noted that it’s important to combine generative AI with other tools to deliver the most powerful value. “What you’ve done now is taken generative AI, combined it with one of the other tools, which is analytical, and reduced the time for information, the time to value for what can be done from weeks to potentially seconds or minutes, and that is super powerful,” Shah said
Streamlining Efficiencies
AI will be a game-changer, especially within the accounts receivable realm. When businesses integrate AI within their AR processes, they will be able to automate the creation of invoices, ensuring timely delivery to their customers. Payments can be processed electronically, and automatic reminders can be sent to overdue accounts.
Billtrust’s latest solution, currently in beta, takes the functions and user experiences of ChatGPT and integrates them in the form of a finance co-pilot within its software-as-a-service (SaaS) application.
“What this is doing is giving you the power of language models on your enterprise data in a secure, compliant way,” Shah said. “This is a private beta, and we believe this is the right avenue to build that interface and that connection with our customers because it’s also new for them. We’d love to understand where we can solve the most pain.”
Handling sensitive customer and financial information through AI necessitates robust security measures that ensure the protection of all users. Equally essential is the ongoing measurement of user outcomes with the launching a new solution.
According to Shah, when it comes to generative AI within the B2B space, meticulous planning, infrastructure development, and engineering expertise are prerequisites. This is particularly evident compared with B2C applications, where the enterprise B2B ecosystem introduces heightened complexity and a substantial volume of data. The data’s cleanliness, organization, and formatting become critical elements, enabling AI models to learn and make precise decisions.
Moreover, integrating generative AI models into an organization’s AR platform requires careful planning and a deep understanding of engineering principles to ensure a seamless flow of data and adherence to compliance factors. When a customer is first loaded into the system, it will have its own segmentation, roles, security, and authentication that will not change.
As for measuring outcomes, Shah says it will be in the form of a “bidirectional feedback loop,” which will include customer counsels and working sessions. He hopes that by tapping directly into what customers need, the company will be able to create the most effective road map for its new product.
Implementing AI in Your Business
AI and its various forms are here to stay, but the question that looms among businesses is whether they should adopt it. As previously mentioned, innovative businesses that embrace and adopt AI solutions will flourish in the areas of efficiency, productivity, and customer experience.
Those that are still on the fence run the risk of falling behind and perhaps stifling their opportunities to scale, especially if they still rely on manual processes.
In exploring the adoption of AI, the best approach is to start organically. Start by embedding it within a few small use cases throughout the organization. From there, test and explore.
“Don’t hesitate to learn and adopt where you can identify very tangible business cases,” Shah said. “Don’t say, ‘I’m going to transform all of my accounts receivable or all of my marketing overnight.’ You need to find a low-hanging fruit.
“What I found is most businesses, if they don’t find low-hanging fruit, they don’t get the momentum needed to actually sustain adoption of a technology. AI is no different.”
