You don’t need big data for AI
Nowadays it seems that every big international company is jumping on the AI bandwagon. That can be a bit intimidating for all of us, with names like Google, Amazon and Facebook using this technology all the time and attributing their success to it. But what about smaller companies? Do they need to follow suit? Or is AI technology out of reach for them? Many SMEs, scale-ups and start-ups are struggling with these questions and are cautiously looking for AI solutions to increase their efficiency, add value to their product and attract customers. Smaller companies often fear that AI is out of reach for them because of a lack of data and resources to optimize processes, at ML2Grow it’s our mission to show them wrong: a little bit of data, can go a long way.
An increasing AI knowledge gap
To avoid being outsmarted by digital competitors, small businesses often use their size to their advantage by using newer and more efficient tech solutions. For example, some of these companies have successfully implemented ERP systems, BI tools, cloud-based software, and other digital solutions in recent years. Yet often it seems that AI is just the one step too far for many small businesses, despite it being one of the most profitable investments out there. Let us be clear, AI is not the future, it is now and companies from every sector are investing in it.
Imec surveyed over a hundred companies in Flanders about AI/ML implementation in their annual AI barometer, which was developed with the help of ML2Grow. Looking at the figures, it is striking that most of the companies did not have any expertise in AI. Around 25% of the companies had some fully developed AI knowledge in their organization. With a majority having only some experience in one or two teams.
Important and worrisome is the link between existing expertise and any plans to further expand the expertise. As shown below, companies without AI competencies often have little or no plans to expand their competencies. Companies that already have AI competencies are more likely to plan to build on this expertise.
Therefore, there is a threat of an increasing AI knowledge gap between the “haves” and “have-nots”. Within the group of companies with AI competencies, we see that 86.2% believe that AI will impact their business. At the companies without AI competencies, only 34.7% expect an impact on their business. But the result remains the same, a growing division. This division between companies that do or do not capitalize on AI has already been examined in international studies. For example, according to an article by the World Economic Forum, it appears that innovative, leading-edge companies that use AI technology can double their cash flow by 2030. This is in contrast to companies that do not adopt AI technology in their operation: they experience a 20% reduction in cash flow.
Bringing down the barriers
Companies see the greatest potential in AI in terms of operational efficiency, followed by customer interaction, discovering new markets and customer segments, product innovation, and business model innovation. International research shows that manufacturing and risk reduction deliver the most significant benefits in AI adoption.
There have indeed been real obstacles to implementing AI solutions for small businesses. Setting up an own data team and internal R&D in AI is costly and the digital-savvy needed to create and train owns model is intimidating.
There are still quite a few concerns within companies about the rollout of AI applications. the main concerns are in the areas of:
- Privacy, security and protection of data
- Transparency and reliability of AI technology
- Job loss
- Reducing social contact
- Limited AI knowledge within companies
- Change management for employees
- Consideration between the added value of deploying AI applications and the cost of the implementation
International research also shows that there are still many barriers for companies to adopt AI. Research by McKinsey identifies a lack of a clear strategy, a lack of competencies and expertise, and working in functional silos as barriers for companies to adopt end-to-end AI solutions. Research by NewVantage Partners (targeting C-level executives) shows that 77% of respondents indicate that the business adoption of AI initiatives remains a challenge for companies. Several factors (organisational, coordination, resistance) are at the basis of this. The respondents indicate that about 95% of the factors are due to cultural challenges and only 5% are related to technology. So if companies want to transform and adopt AI initiatives, they first have to start getting rid of cultural obstacles.
For example, it happens frequently that money and time are invested by management in an AI system that is not (fully) used by others within the organization for whom it provides added value. What is unknown also remains unloved, and many employees have a negative view of AI. This often finds its origin in misuse of the technology that is covered in the media. There is also a persistent myth that an AI system can replace employees and their knowledge, making the technology a threat rather than an opportunity.
In our previous blog post ‘how to make AI successful in your business‘ we also stressed the point that AI is above all a cultural change. The most important change has to take place in the culture of the company. Interdepartmental collaboration and information sharing must be encouraged to make the adoption of AI successful.
AI is like electricity, as coined by Andrew Ng.
We fully embrace this at ML2Grow, AI will be omnipresent, empowering solutions in the background. – Joeri Ruyssinck
Why novel breakthroughs in ML are gamechanging
Belgium is know for its SME focused economy. Because of their size, they have typically less data available than large companies. It is known that AI/ML applications become more effective as they are fed with more data, so this could create a growing barrier and competitive disadvantage for SMEs to play the AI game.
Thankfully, the concerns and costs of small business AI integration are now mostly obsolete with recent breakthroughs in ML. Progress in data-efficient machine learning, infrastructure and the availability of pre-trained models in deep learning are making powerful AI available with less data.
Deep learning is actually a type of machine learning that can process a wider range of data resources, requires fewer data preprocessing by humans, and can often produce more accurate results than traditional machine-learning approaches (more on that further in this post). The big downside is that it requires a much larger amount of data to do so, or is it?
Luckily, this is not always an issue anymore because pre-trained ML models are helping small businesses to build solutions that used to only be available to large enterprises. Because ML technology has been in development for a couple of decades now, you don’t have to start from scratch. Many pre-configured networks are available and we can successfully adapt and refine them with less data to the specific solution in mind.
In deep learning, interconnected layers of calculators known as “neurons” form a neural network. The network can ingest vast amounts of input data and process them through multiple layers that learn increasingly complex features of the data at each layer. The network can then make a determination about the data, learn if its determination is correct, and use what it has learned to make determinations about new data. For example, once it learns what an object looks like, it can recognize the object in a new image. More on this can be found in one of our previous blog posts.
Especially in computer vision (image classification, facial recognition, maintenance prediction, automated quality control) and NLP (voice recognition, chatbots, automated call centre) did the cost of AI adoption significantly decrease.
Open-source frameworks have become widely available. The ROI is high because these solutions save you time on mundane tasks and turn data into insight. Integrating AI into business processes can help ensure more timely and more accurate decision-making.
Use cases and the benefits for your small business
Improve hiring and retention
Hiring and retaining talent is a huge responsibility for any business; however, smaller companies often have either no formal HR department or a small team or solo practitioner. Even if your business is small, the amount of human resources work can be daunting for a single individual.
Thankfully, HR technology has evolved to include AI and machine learning to reduce repetitive tasks and speed up functions that used to take a lot of manual work.
In our blog post ‘Why AI is needed in Human Resources‘ we go deeper into the importance of AI in HR.
Small companies can make a significant impact by using online marketing solutions enabled by machine learning and AI. Modern consumers are looking for personalized, relevant customer experiences, and AI can help small companies deliver that.
A powerful way to boost your sales, customer satisfaction and customer loyalty is to use a recommender system. Such a system enables you to draw up user profiles based on their behaviour and to predict which products are most relevant to such profiles. A user profile specifies the properties of how much interest a customer has in a particular product.
This information can be used to target your marketing to the most optimal target audience at the right time.
In our blog post ‘how to boost your sales with a recommender system‘ we explain why a recommender system is a valuable asset to maximize your marketing investments.
Improve customer experiences
NLP tools are helping companies understand how their customers perceive them across all channels of communication, whether emails, product reviews, social media posts, surveys, and more.
You don’t have to increase headcount to serve your growing number of customers. With the implementation of an AI chatbot, your website can become a first stop for solving problems and answering customer questions quickly.
In our blog post’ five applications that show the power of NLP for your business‘ we show the practical side of NLP in business.
Increase sales efficiencies
Some research shows that it costs five times as much to acquire a new customer than to keep an existing one. What’s more, when nurtured correctly, existing customers are more likely to keep purchasing and become brand loyalists. Brand-loyal customers tell their friends about your company, are more likely to try new products, and are less likely to be chased away by a price increase.
Customer churn can manifest in different ways for different businesses, but it can be generally defined as what happens when a customer unsubscribes from your service, ceases to purchase from you, or simply stops engaging with your brand. Customers leave for a variety of reasons, and it’s an inevitable part of doing business.
Still, too many companies fail to properly address customer churn. While usually no expense is spared to find and attract new customers, the same seldom applies to identifying customers which are ready to join your competitor. But how exactly can you focus your efforts on the right customer?
In our blog post ‘our secret to predict customer churn‘ we explain how you can easily identify and target customers with personalized offers.
Help improve decision-making and processes
A lot of the AI buzz you hear centers on advertising, marketing, sales, and even HR. However, small companies can use machine learning to improve decision making and processes in multiple areas.
In a previous blog post ‘how you can automate inspection tasks with modern computer vision‘ we show how computer vision can have a massive impact on companies of all industries. Sensors can perform maintenance prediction to automated quality inspection in a production line.