Taken from Forbes
Written by Laura Rocha, Co-founder and CEO of Dathic
99.9% of US businesses are still far from being able to develop their own Artificial intelligence (AI) models and lagging behind in implementing AI-powered operations. Training AI models requires access to big amounts of data and computational power, a high-tech skilled workforce, and a lot of upfront cash flow, which is beyond the means of small businesses (SMBs). As this business segment accounts for 99.9% of the US businesses1, there is a huge need and opportunity to make AI less “artificial” and extraneous and more assistive to drive the growth of more than 33.2 million small businesses in America.
Since 1955, emeritus Stanford Professor John McCarthy defined Artificial Intelligence (AI) as “the science and engineering of making intelligent machines2.” Although it might seem like a new trend that computers can generate an output like a text or an image, it has taken many decades to grow computational power and data to achieve the current AI capabilities and levels of impact.
The development of artificial intelligence algorithms and AI-based solutions has been a very capital-intensive, risky, and long journey led by a few companies. After decades of research and billions of dollars deployed on testing and tuning AI models, large companies are already enjoying the benefits AI brings in efficiency and competitive advantage that are almost unattainable to the smallest businesses.
The idea of training algorithms to understand, identify patterns, classify, or generate outputs came mostly to solve complex and large-scale problems in large corporations. In fact, one of the foundational applications was OCR (Optical Character Recognition), it gave computers the ability to read handwritten images, like checks, and convert them into text to be easily processed by computers. This was a very common problem for large companies, including banks back in the late 80’s. Today, AI solutions are still at a high complex level not suitable for SMBs; but as this segment represents the majority of businesses, there is a huge opportunity to reframe the way Artificial Intelligence is developed making it “Assistive” Intelligence and integrating AI in the day-to-day business decisions of SMBs.
I heard of this concept of Assistive Intelligence back in 2019 at the KDD, the annual Conference on Knowledge Discovery and Data Mining, and I've come to this photo on my phone so many times as I keep working on AI applications for small and mid-sized consumer brands.
At this conference, the conversation revolved around how just a few businesses will have the highly trained workforce and technology capacities to develop the whole power of AI algorithms, and how most of the businesses will be adopters with restricted or limited technical capabilities.
Assistive design has been commonly applied to develop infrastructure, such as sidewalks, and train stations friendly and accessible to people with limited motor or physical function; allowing them to move or perform their day-to-day activities in an independent manner. The idea of Assistive Intelligence implements a similar train of thought, making rather complex technology more accessible to businesses with limited technical resources. It will be very important, sooner rather than later, to design AI-powered systems strategically and intentionally to assist Small businesses and allow them to interact independently with AI-powered solutions and experience the benefits in efficiency, as they would not be able to develop it by themselves in the near future.
Why focus on small businesses when it comes to AI adoption?
SMBs are companies with 500 or fewer employees. With over 30 million SMBs in the US, it's the largest and fastest-growing business segment, which accounts for 99.9% of all U.S. businesses. They also have a huge influence in the labor market, as they employ almost half (46%) of America's private sector workforce, according to the Small Business Administration (SBA).3
SMBs represent 43.5% of the US gross domestic product but just 32.6% of known export value. Al-powered processes can help SMBs to better forecast, and optimize time, materials, or logistics. This can make them more competitive not only in their local ecosystems but also expanding their presence in overseas markets, ultimately powering the growth of the country’s economy. In China, small businesses account for a similar percentage of 98.5% of total businesses and represent 60% of the total GDP and more than half of the country’s exports (68%)4.
AI can help SMBs work smarter and not just harder. In my work with founders in the Consumer Packaged Goods (CPG) space, I hear how challenging and stressful it is to make critical decisions due to the drastic consequences if things go wrong. They are constantly assessing different scenarios to determine how to allocate their limited resources and most of the time feel insecure about their strategies or have to go blind as they lack information to guide their decisions.
Producing, marketing, and selling a product is a complex process for small businesses and AI can provide an extra set of hands when making growth decisions. Empowering SMBs owners with data and AI-powered technology can take the guesswork out and optimize their time, resources and grow their sales. A common challenge Dathic helps CPG brands with is finding the best stores where to serve their multicultural consumers. As humans, we usually consider no more than five variables to make a decision; Dathic’s AI-powered models analyze over 100 variables to recommend CPG brands the best selections.
Placing a product in the wrong store might generate low performance, demand higher marketing expenditure, and even bans from retailers. With limited and extremely competitive shelf space, brands need to outperform retailers’ volume and velocity expectations to keep growing in the store. Adopting a data-driven strategy to select stores, can lead to resulting growth of more than 25% in sales as well as growth in customer loyalty by serving local communities more authentically.
As if there is still doubt about the gigantic market opportunity and the economic benefits of developing AI to assist SMBs; There is a proven record that this segment is eager to adopt tech and when experiencing the benefits, they are loyal and profitable clients. In 2020, Small businesses account for nearly 75% of Facebook’s $70 billion annual ad revenue, according to Deutsche Bank, and according to Facebook reports in 2019 the 100 largest ads clients account for less than 20% of the ads revenue5.
How can we make AI Assistive for SMBs?
There is already a digitalization gap among SMBs. Most SMBs don’t have clean databases, robust tech infrastructure, data scientists, or business intelligence teams. Often I hear business owners say: “I know, we’re so behind, but we want to invest and learn to handle the data and use the technology to grow.”
There is no reason to wait until they reach the operational (and financial) levels of large corporations to introduce data-driven decisions or processes powered by AI. Instead, it would be more effective to introduce SMBs into AI sooner than later. It's simpler to implement in small processes, cheaper to set up a lighter infrastructure, and mostly it’s easier to create a new cultural mindset in a small team.
SMBs need AI solutions that are intentionally designed for them, not just mini versions of what's made for larger companies. These solutions should be actionable, explainable, and delivered with empathy.
Actionability is key for SMBs to consider before investing in new solutions. Business owners expect a tool they can confidently use with little training and get results in the short term. Recently I heard an AI supplier said that it takes on average 50 weeks to experience the real power of its tool. SMBs often cannot afford to have a dedicated staff for extended training periods or wait for the multiple implementation phases that are typical in introducing AI solutions to larger corporations.
Tools that “translate” technical outputs into digestible insights for their specific business context are a dream for SMBs. The explainability of AI models is a big challenge; being trained with billions of data points, and using a complexity of structures it is not always easy to determine the reasoning behind their outputs. As a lawyer, coming from a social science background, I understand the relevance of making it easier for small and medium-sized businesses without highly skilled technical staff to process data and translate outputs. By providing insights in clear language, I hear my clients say how valuable is for them to save up to 90% of their analytics time, so they can make their daily business decisions more easily.
The fact that 99.9% of US businesses are SMBs, less than 2% of them have more than 20 employees and on average, an early-stage startup has 4 employees needs to be at the top of the mind of AI developers and providers. The AI race is currently led by a few companies in the private sector with a minimum representation of the small business segment, so creating solutions that really assist SMBs and delivering to them with empathy must be a priority. Investing in SMB efficiency can power our economy, create a stronger business ecosystem and improve consumers’ life.
2 Stanford University - AI definitions
Follow me on LinkedIn.
