The Biggest Obstacles to AI Adoption for Fintech Startups
Every company out there wants a piece of artificial intelligence. From large tech companies like Google and Amazon to small businesses trying to survive in a highly competitive business world, the benefits – both real and perceived – are too enticing to ignore.
According to one report by Vanson Bourne, about 80 percent of companies are actively using one or more elements of AI, including machine learning and natural language processing. In the same report, a further 36 percent of senior decision-makers said they would be making more AI investments within the next three years, with at least 62 percent hoping to bring in an AI officer in future.
These figures are hardly a surprise. AI has been the main driving force behind most fintech disruptions, which makes it a game-changer for small businesses and startups that are coming up within this field.
Although businesses are taking steps to bring in AI into their companies, the AI playfield is fraught with obstacles that keep the best of AI away from most small businesses. From team members who are clueless about the technology to legacy systems that won’t support the tech, things can go south quickly for business owners who are not prepared.
If you’re planning to bring in AI-assisted technologies for your business in any form, here are some of the major issues to keep in mind before jumping in.
Legacy IT Infrastructure
In the Vanson Bourne report, close to half of the respondents identified the lack of an IT infrastructure as the biggest obstacle to AI adoption. Keep in mind that the respondents were top-level managers in companies with $50M-plus in revenue annually.
AI systems, even the most basic ones, run on resource-intensive algorithms that require a decent combination of hardware and software capabilities, something that isn’t usually available in your typical office setup from a decade ago. At its basic level, an AI-ready infrastructure should be capable of efficient data management, have enough processing power, be agile, flexible and scalable, and have the capacity to accommodate different types and volumes of data.
Naturally, it would be more challenging for fintech startups and small businesses to assemble the necessary hardware and software elements to support AI. Still, with proper planning and subtle changes to your legacy systems, you can make AI part of your strategic future. Instead of bringing in new servers with powerful processors, for instance, start with a cloud-based service and move on from there.
Problems with Big Data
Data is the driving force behind all intelligent systems. And with the amount of data crossing the 20 zettabytes threshold last year, quality data, and not smart machines, has become the most important prerequisite for developing AI systems.
For fintech startups, two issues present the biggest obstacles to creating or adopting data-driven AI solutions. First, there’s a large amount of unstructured data on the web – up to 90 percent of internet data according to some reports. Unstructured data is any pool of poor-quality data that cannot be understood by an intelligent system, which greatly limits what an AI system can do.
Secondly, there’s always the sensitive issue of data privacy and security. In a social experiment highlighted by mSpy, the research team intentionally lost smartphones to help them gauge public reaction to lost phones. Over 50 percent of those who picked up these “lost” phones tried to access the personal data and information on them, illustrating just how tricky data security is in the real world.
More recently, the Facebook data scandal that happened a few weeks ago served as the perfect reminder of just how fluid privacy issues are when it comes to handling data.
The Future of AI for Fintech
What do these issues mean for fintech startups? For one, as regulators become stricter with how businesses handle data, they introduce additional overheads for small businesses that must scrutinize every byte of data.
Businesses can’t simply harvest data from the internet or CRM databases for their AI solutions, but will have to spend time and resources going through each data set to make sure it’s AI-digestible and doesn’t violate any regulations. Thus, in addition to the lack of talent and managerial skills required for setting up AI systems, startups might have to take up the services of professionals like data scientists, something that’s often out of reach for startups.
In the end, however, there’s always hope for businesses that want to put in the effort to bring in AI. Understanding the challenges ahead is an important first step for enterprises with a focus on the future.