SMEs adopt new technologies not just for the perceived impact on their bottom lines but also because of the fear of missing out on the competitive edge that their competitors might gain. Even so, most SMEs are sluggards in embracing new technologies.

A recent study by McKinsey revealed that SMEs who are early adopters of technology will double their cash flow by 2030 as opposed to those who don’t. But it is not just the loss of revenue that SMEs should be worried about.

The fifth industrial revolution is here with us and with it comes very unique challenges for small enterprises. Once the big fish attain their 5th industrial revolution goals, it will be an uphill task for the SMEs to remain afloat. The question begs – is there any hope for SMEs in the fifth industrial revolution? Could the democratization of AI offer them some reprieve?

Will SMEs stand a chance against AI-powered big-corps?

Huge corporations like Amazon, Netflix, Tencent, etc. are already reaping the benefits of AI-powered business models e.g. by creating personalized offers which continually evolve with respect to their clients’ likes and preferences. As the agile corporations take advantage of various AI technologies to better their business models, some SMEs are making the expensive mistake of maintaining the status quo.

This is the same mistake Lotus made in the 80s. Back then, Lotus’ product had an enviable market dominance that can only be compared to today’s Google. Lotus was actually the primary reason why people bought PCs back then.

Around the same time, Windows started developing excel, a product that was meant to compete with Lotus 1-2-3. Instead of responding positively to this new challenge, Lotus decided to ignore windows. They figured people would soon get over windows. To their surprise, most of their clients switched to windows as soon as Microsoft excel was deployed. Since they were bleeding customers, Lotus had to rush to try to add some backward compatibility to their product but it was too late. Eventually, Lotus lost their business completely and they were eventually bought off by IBM. That is how a great Lotus product became extinct.

One important lesson every SME can glean from the likes of Lotus is that technology can be mercilessly disruptive. Those who are slow in adapting to new technology will pay a dear price while the early adopters will reap the benefits.

AI tools can actually help the SMEs to create hugely successful enterprises that will compete relatively well with the big multinationals. For this to happen, the SMEs must know how best to leverage various AI tools that are within their reach.

Most small enterprises shy away from machine learning and other innovative technologies because of the misguided notion that they will have to break the bank for successful implementation. But it is actually possible to launch technology on a smaller scale by leveraging ready-made solutions. The following are some suggestions on how SMEs can implement AI on the cheap.

Leverage existing solutions

The huge companies have the financial muscle to develop their own complex AI solutions but you don’t have to. There are lots of tech companies that have come up with open-source projects which you can use to save your company money and time. A good example of such a platform is Google’s TensorFlow which was designed to encourage research and production. Some of the big names that use TensorFlow include Twitter, Uber, Dropbox, Intel and eBay.

Start small

Rome was built but not in a day. It is never a good idea to rush all-in into any new venture. As an SME, it is safest to use a gradual adoption approach. Start small and grow your AI efforts progressively. For instance, you may want to start with the integration of third party apps that foster productivity of your workforce.

Once that has been achieved, you can move on to open-sourced AI and cloud systems. Focus on solving problems that will give you the highest ROI in the short term.

AI for analysis

Predictive analysis can help you scale up the operations of your business in unimaginable ways. Analytics is one of the safest and quickest ways for an SME to implement AI without having to spend lots of money on Machine Learning. Some of the common products you can use for this include the Amazon Machine learning, Microsoft Onboard, and H20

Are we moving towards mega-companies and the singularity?

Kurzweil Ray, the director of engineering at Google publicly stated that he believed the singularity will be realized by 2029. Granted, experts may not agree on the timelines but they agree that the singularity, a phase when machines will have human intelligence is inevitable. In fact, emerj surveyed 32 PhD scholars in the field of AI and 45% of them agreed that it was very likely to achieve the singularity before 2060.

As technology continues to advance, we can expect computers to merge with nanotechnology, robotics, genomics and other technologies that will make progress in tech virtually instantaneous. This notion might have appeared far-fetched couple of decades ago but recent advances in AI would make anyone think twice before taking the issue lightly.

IBM’s Watson is already helping their staff to fight cancer while at the same time helping in other process like cooking and financial planning. Automakers are in a race to release their versions of self-driving cars by 2020.  These two examples are enough to show that technology is moving pretty fast.

Computers are advancing quicker than any other man-made device. Computers were created to help humans accomplish complex tasks faster and with greater accuracy. Likewise, intelligent systems are created to perform intelligent tasks autonomously or with very little input from humans.

Economic pressure is already forcing the big corporations to develop systems that will make their process more efficient and profitable. For instance, the work that was accomplished by hundreds of employees a couple of decades ago can now be achieved by one industrial robot and at a fraction of the cost too.

The benefits of intelligent machines clearly outweigh any of its demerits, especially for the big companies. This will continue mounting pressure on companies to continue developing better AI and this, experts believe, is what will fasten the achievement of the singularity.

The AI/ML problem is really a data problem

AI is increasingly becoming inherent in how we operate. But AI is not a flawless solution to all the problems we might have. Every AI is a product of an algorithm and quite unfortunately, most of these algorithms inadvertently inherit some biases that might end up perpetuating some of the challenges the AI was designed to solve in the first place.

For instance, Amazon’s recruitment system was designed to analyze prospective employees based on the data provided and then recommend the best candidates from the applicants list. The AI would even notice some important qualities and traits that the HR professionals would probably miss.

Unfortunately, the system ended up rating male candidates higher than female candidates and Amazon had no choice but to discontinue the system. Another great example is Microsoft chatbot.  Within 24 hours of its launch, the chatbot had been taught by the twitter users to post racist and demeaning remarks. Microsoft had to pull it down to avoid further damage.

Both of these examples have one thing in common – the real problem is not the AI but the data. According to the World Economic Forum, there is a need to address these systemic issues in data science in order to avoid grievous mistakes in future AI technologies.

The biggest problem in most AI systems today is that they are designed to detect patterns in data using a top-down approach. Because these AI systems are largely dependent on deep learning, they easily reach a plateau in certain use cases and this means the AI will not be able to explain the solution arrived at.

Human countermeasures need to be used to deal with these biases. Techniques like use of more diverse data sets and increasing diversity in a given industry might help to deal with the biases.

Leveling the playing field

The Hype Cycle for Emerging Technologies report which was published by Gartner in 2018 identified a couple of emerging technologies that are bound to blur the lines between machines and humans. These include Ubiquitous infrastructure, DIY Biohacking, Digitized Ecosystems and Democratization of AI. But what makes AI an interesting trend is the fact that it will probably be used by everyone.  In other words, AI will be democratized.

According to Gartner, democratization of AI will actually benefit companies by helping them to considerably cut down costs. For instance, a worker that uses self-service analytics will probably give more output in terms of analysis as opposed to a professional data scientist.

And it is not just mere analytics that can be made accessible to the employees. Some large computing providers have already created open source tools that novices can take advantage of and build their machine learning models. These tools come with prebuilt algorithms that make it easy for a user with very limited experience to begin.

Google recently released Cloud AutoML which uses machine learning to automatically build and tweak Neural Network for image recognition. Another prime example is DataRobot, from DataRobot Inc, which allows a user to upload his or her data, from which the system will build lots of models.

There are lots of open source ML libraries that are coming up each day targeting budding developers. Granted, you will still need some coding experience to use them but the fact that they provide the basic sub-components that one would need to craft their custom algorithm can help novice coders to develop powerful AI applications.

Is the loss of SMEs a future fact or just a bleak outlook

While the big companies embrace technology as their path to gaining competitive advantage, most SMEs shy away from it because they view it as another cost center. Unfortunately, the digital transformation might easily wipe out any SME that refuses to embrace change.

Digital transformation isn’t only meant to deliver incremental efficiency. It is meant to completely transform the way of doing business. There are a number of examples from history that should serve as a warning for those that are resistant to change.  When the technology of photography changed, Kodak didn’t ride the tide with the rest of the people. Most retail companies didn’t take Amazon too seriously during its infancy and Cab companies probably had no idea that Uber would completely take over their business.

The worst mistake any SME can make is to think that this kind of disruption only happens in certain sectors or certain countries. Most of the business dinosaurs that became extinct were probably hit hard with a technological shift that they were not prepared for.

They were comfortable with business as usual and before they knew it, their market share had been swept from right under their feet by other forward thinking startups. Every SME should start gradually adapting to the technological shift before change suddenly catches up with them.

It is time to get this right before you get left behind.