90% of businesses understand the need to enable data-driven decision-making at all levels, but just 7% are taking steps to make that happen. Why the disparity?
COVID-19 has changed much about the world in just a short time. In the business environment, specifically, those changes include an acceleration of automation and digitalization — and even an increase in the amount of data that organizations are generating and storing. Documents and images are now the key tools businesses are using to interact with their customers.
Most organizations aren’t able to harness the value of these and other forms of unstructured data, however. And, as I explained in a previous post, the business problem of managing data within documents and images will continue to be a struggle for many companies without the right tools — namely, artificial intelligence (AI). AI is the key to helping them understand the data that is trapped within their document and image files and to ask meaningful questions about that data to elicit business insights.
Additionally, by adopting AI, businesses can empower more people in the organization to contribute to data-driven decision-making. It’s a smart move that can deliver a solid return: McKinsey research finds that high-performing organizations are three times more likely than others to report that their data and analytics initiatives have contributed at least 20% to earnings before interest and taxes (EBIT).
Yet, according to Harvard Business Review Analytics Services research, while 90 percent of businesses understand the need to enable data-driven decision-making at all levels, just 7 percent are fully equipping their teams with the tools to make that happen.
Why the disparity? Well, think about where many firms were with AI before the pandemic. That is, in most cases, nowhere. In fact, research from Protiviti and ESI ThoughtLab found that less than 20 percent of companies around the globe were gaining significant value from AI. Notably, that number was expected to increase threefold by 2021 as more businesses fast-tracked AI investments — but that was before COVID-19.
The study also noted that — not surprisingly — technology companies were among the businesses moving the fastest with AI. But while they were leading, they were also lagging in some ways. According to separate research by Bain & Company, nearly 90 percent of tech executives, prior to the crisis, were not satisfied with their company’s approach to AI. Still, 90 percent of these leaders said they see AI and machine learning (ML) as priorities to incorporate into their product lines and businesses.
Today, many organizations remain in a state of AI inertia. Some worry that now is not the right time to make “risky” investments in emerging technologies. Others have yet to feel a sense of urgency about AI, continuing to categorize it as “something we might explore in the future.” But in most cases, companies aren’t using AI because they don’t know how, where, or why to do it. That was a big obstacle to AI adoption before the pandemic — and it persists.
That brings us to practical AI, or putting AI into action to solve a specific business problem. For AI to be considered “practical,” you need a clearly scoped use case with quantifiable ROI (e.g., time saved, revenue generated), and the right setup for AI and ML to be applied successfully with the inclusion of a well-designed feedback loop. (This loop helps drive continuous improvement in AI projects and prevent their failure, as this post explains.)
An example of practical AI already in use in a growing number of organizations is simplifying the process of surfacing and organizing data assets, including unstructured data like images and video, for use on eCommerce channels. With an AI-powered, searchable content repository for product data, imagery, and copy, they can deliver and display accurate and complete product information to eCommerce channels faster — using fewer resources.
We’ve also seen growing interest in use cases such as automatically extracting data from invoices and bills to analyze spending patterns and streamlining the mortgage loan process by processing applications automatically. There are literally hundreds of applications for practical AI.
Practical AI can also help organizations make the best use of their talent — including remote workers. Employees’ access to available and complete data is a must for remote workforce efficiency because it increases team productivity and reduces costs. Some companies that are utilizing practical AI solutions are saving seven figures per year.
Automating data-intensive tasks also frees top performers to focus on higher-order work, which is critical for businesses that want to thrive in a post-COVID-19 environment. In particular, companies will want to unleash the full potential of the innovators in their workforce, who can help the business to scale. These tech-savvy workers in marketing, sales, human resources, finance, and other functions constantly seek out and evaluate better solutions and are open to trying new things for the sake of greater efficiency and more productivity.
The innovators will expect the company to provide more visibility and access to data across the business to create new ways of working and productivity breakthroughs. But without AI, organizations can’t come close to meeting that expectation. They won’t be able to automate tedious, data-related tasks, and empower their workers to surface, find, and share assets faster and more effectively. So, their employees will continue to spend valuable time and energy searching for and organizing data instead of actually putting that data to work for the business.
The burden of manual processes will also prevent employees from engaging in meaningful work — for themselves and for the company. That could negatively impact not only workforce productivity, but also retention. Consider that 41 percent of employees are already feeling burned out, drained, or exhausted by work during the pandemic, according to an SHRM study.
While investing in AI may be uncomfortable in a time of disruption, procrastinating could prove even more detrimental for businesses in the long term. Practical AI provides a clear answer to the “how” of getting started with the technology and also addresses the questions of “where” and “why.” Pursuing practical AI projects can provide a shorter path to positive results while also laying the groundwork for an AI-powered, data-driven future for the business. And many industries, from automotive to energy to travel and hospitality, are ripe with use cases for practical AI.
Take the financial services industry, for example. This is an industry that deals with mountains of paper statements, and the process of extracting data from documents like invoices is time-consuming and costly. Many firms look to outside parties to handle this work because their employees simply do not have the bandwidth and the company wants them focused on higher-value tasks. However, quality control issues often arise with this approach. And if the business needs to scale this work — for example, because it’s launching a new product that requires even more processing of paper statements for clients — costs can quickly escalate.
The answer is to keep the process in-house — without undermining employee productivity. Using traditional document processing technology like optical character recognition (OCR) isn’t the answer, however. Template-based OCR solutions can extract data, but they don’t provide context (and context matters, as this post explains). Also, any text to be extracted must appear within the boxes defined by the OCR template. So, if invoices for processing don’t conform to the template, the template essentially breaks, leading to unreliable and inaccurate results.
Scaling this process is also impossible — unless you bring in tools that can learn and adapt to variability in documents. Impira’s AI-powered automation platform can do that. It can extract totals, invoice numbers, payment terms, due dates, and more from invoices across vendors and different formats in just a few clicks. And users can then export data directly into a spreadsheet.
It’s also easy for non-technical users to build ML models in Impira that can understand and create structure from their documents. Simple interactions, such as highlighting a word, labeling, or clicking a checkmark, become real-time feedback (i.e., the all-important feedback loop) that updates the ML model and returns predicted text right back into the corresponding cell.
That’s a quick look at practical AI in action. It’s applying AI and ML to solve a real business problem, deliver quantifiable ROI (saved time, more efficiency, scalability), and give users the ability to drive continuous improvement. Is your business ready to get started with practical AI? There’s no more time to waste. Archaic, manual processes prevent employees from doing their best work. And without AI and ML, businesses and their teams simply can’t manage, understand, or effectively use their vast and ever-growing stores of data.