the cognistx blog

Measuring the ROI of AI

May 13, 2024

By now, most executives and business owners know investing in AI solutions is essential to remain competitive and abreast of technology trends. Yet the investment, while necessary, often leads to questions about calculating the return on investment (ROI).

At Cognistx, we place a high value on customer feedback and actively incorporate it into our product development, particularly with SQUARE.

"We hear from a lot of business owners that are actually making these decisions based on expectations of very immediate return on investment for some of these AI solutions. And I think that's one big misconception is that not all AI solutions are going to be an immediate source of money into your company," said Cognistx Data Scientist Justin Waltrip.

Many times, AI solutions represent long-term investments that require time, as well as funding the  technology, people, and processes. Recently, there has been a surge in companies promising immediate ROI, suggesting that using certain tools will instantly revolutionize company operations.

"That's not how these solutions work. They take time and they take iteration and feedback in order to really customize it to your particular business case.and make sure that the tool is operating exactly as it should be," Waltrip said.

Managing Client Expectations

Businesses need to understand the basics of technology to know its strengths and weaknesses. This knowledge keeps them from having unrealistic expectations about what AI models can and can't do.

At Cognistx, tackling the challenges AI can solve is our top priority. However, it's just as crucial that users fully benefit from the technology. For new clients, especially those skeptical about AI's impact on ROI, our first step is to educate them on what AI can realistically do. This way, we can set clear and achievable goals for them.

It's key to really get what problems users are facing to make the most of AI solutions. We should be solving real issues, not just coming up with solutions looking for a problem. By talking to clients and understanding what they need, we can create AI tools that actually make a difference and offer real benefits.

"It's essential to move beyond the notion of AI as a one-size-fits-all fix, focusing instead on how it can solve real-world problems in a way that enhances employee adoption and overall effectiveness. By prioritizing the actual needs of users from the outset, we can ensure that our AI solutions are not only innovative but also genuinely beneficial," Waltrip said.

Making Adjustments Based on Customer Feedback

Client feedback is essential. There are going to be issues, and the most important way to address them is by collecting that feedback from the clients and using it to potentially address those problems.

At Cognistx, we've seen how important it is to add flexibility and customization to our platform SQUARE, our AI-powered enterprise question answering tool. We let and encourage potential customers  use the product, much like you would test drive a car. As they ask questions and SQUARE provides an answer, we collect thumbs up and thumbs down feedback from them.

Feedback is also built into our newly released AI-website assistant SQUARY. Users let SQUARY know if their questions provided the information they wanted or if it didn't. Because SQUARY, learns from user engagement and feedback, improving over time.

We also seek out more subjective feedback from clients. They might express preferences like disliking the tone or wishing for references to specific documents in the responses. This kind of feedback is invaluable because it allows us to fine-tune the model's behavior specifically to meet each client's needs, addressing their concerns more effectively. This approach is particularly important as we continue to evolve our AI tools.

"We're aware that there are going to be business restrictions in the way that AI can be used. We can't have a chatbot that says crazy things or says things against the company policies or answers in a tone that just doesn't really align with that company. That's where that feedback comes in to play. And we really need to listen to the client and think about how we can actually address their concerns," said Sanjay Chopra, CEO of Cognistx.

How should companies calculate the ROI when it comes to an AI business tool

When considering the integration of AI solutions into your business processes, it's essential to evaluate the specific issues you're aiming to resolve. Focusing on the productivity benefits is a useful approach, given the significant potential return on investment (ROI) AI technologies can offer. Conducting comparative assessments of AI tools is crucial to making informed decisions.

Take, for example, a legal firm burdened with the weekly task of extracting agreement titles, involved parties, and specific termination clauses. This manual review process is not only time-consuming for its staff.

Incorporating AI into these processes could dramatically enhance efficiency and accuracy, reducing the time and cost associated with manual reviews. The key is to identify and implement AI solutions that align with a company's unique business needs, maximizing productivity and overall ROI.

"Integrating artificial intelligence into business processes isn't about replacing employees; it's about enhancing the efficiency of their existing tasks and workflows," Waltrip said. "By adopting AI solutions, employees can review documents more swiftly or handle more documents within the same time frame. This efficiency boost allows them to dedicate more time to client communication or engage in higher-level strategic thinking, such as synthesizing vast amounts of information. "

Evaluating Savings

When evaluating the ROI of AI, it's crucial to consider the cost savings from time reductions in tasks like tedious information extraction work, alongside the value of enabling employees to perform more complex tasks.

Companies are beginning to recognize the tremendous benefits of incorporating advanced technology into their operations. Not long ago, the proposition of integrating advanced AI workflows to achieve marginal gains, such as a 10% time saving, was a challenging sell, Chopra said. However, the landscape has dramatically changed with the recent advancements in language models and AI technologies, he said.

Today, the efficiency gains from using cutting-edge technologies can lead to significant time savings, often exceeding 50%. This leap in performance is not about replacing the human touch in business processes but enhancing the ability to analyze and manage information more effectively. For instance, in tasks requiring the review of thousands of pages to identify specific contract provisions, AI can swiftly locate the necessary details, saving hours of time that could be spent on other tasks. AI can summarize and extract key information, provide examples for fast human review, and help compile detailed reports more easily.

Additionally, the significant enhancements in AI models in recent years have shown businesses the potential for notable time savings. This realization has made it easier to see the return on investment these technologies provide, proving why investing in such advancements is not only beneficial but necessary.

Intangible benefits of AI

Exploring how AI can benefit a company beyond just increasing revenue is important. Many people see the value in this, even though the benefits can be hard to measure and explain. In general, there are several key benefits to think about. For example, one major advantage is improving user experience, which is especially clear in the case of Square.

At Cognistx, our goal is to make employees more productive by helping to automate some of the more tedious parts of the work that they would do every day, Waltrip said.

"No one enjoys opening up a hundred page HR document and trying to control F to find the specific language that they might need to reference. And if they don't know exactly where it is, maybe reading through different parts of it to find exactly the information they're looking for. That's not generally a process that your employees actually enjoy. And so if we can help to automate that process by having the AI take over some of those components, then we could potentially improve that employee's experience during the day and help them focus on some of the more interesting and high level aspects of the job that they actually do,"  he said.

Improving employee satisfaction can be hard to measure, but the long-term benefits are clear. One major benefit is better employee retention. Keeping up with technology not only helps keep your current staff but also draws in experienced professionals. This boosts your tech team's skills significantly, giving you an important, if not always obvious, edge.

The same idea holds for customer interactions, especially with customer-facing AI technologies. We can definitely improve the customer experience. But, it's really important to be careful when we start using these systems. Being clear about using AI and making sure it actually makes the customer experience better, not worse, is key.

Dropping Cost of AI
As AI technologies become standard in the industry, economies of scale will kick in, cutting down the development and implementation costs significantly.  With technology getting easier to access and continuous improvements in AI frameworks, tools, and cloud services, it's evident that these technologies will become more affordable.

"From the vantage point of someone deeply involved in the daily implementation of these systems, I've witnessed a remarkable improvement in the tools for building AI systems over the past two years. This advancement means that the overall cost of these solutions is set to decrease, primarily because the need to build everything from scratch is eliminated. There's no longer a stringent requirement for a large team of data scientists, which is a game-changer," Waltrip said.

Listen to AI-Driven for more on the ROI of AI.

Past Blog Posts