AI is being likened to a “Fourth Industrial Revolution” because of it’s potential for creating profound change in people’s lives, much like the invention of steam engines (First Industrial Revolution), electricity and mass production (Second Industrial Revolution), and the rise of the digital age (Third Industrial Revolution) had in the past.

AI/ML applications are everywhere, from ride sharing companies to voice search on mobile phones. We use AI every day – it feels so accessible to consumers that it must be within reach for small business leaders, too.

Given the inevitability of AI’s applications across sectors, it would seem a good idea for small businesses to jump on the AI bandwagon sooner than later to be more competitive… Right?

Not necessarily – timing will be critical for small business adoption – and in this article I’ll aim to lay out a framework for how small businesses should prioritize the adoption of AI in the years ahead. This article will explore how small businesses can use artificial intelligence now, and you will see why it is very important that small business leaders act with prudence.   

How Small Business Leaders Should Think About AI Now

Despite the seeming proliferation of artificial intelligence technology today, it is still in the very early stages of development. Contrary to what might be hyped on social media, in nearly all application areas AI is an expensive and complex solution without evidence of direct ROI. This is especially relevant for small businesses with limited data, limited resources, and limited data science talent.

Modern machine learning is good at detecting patterns in data. It is good at optimizing processes in some cases. By no means does AI solve nor do whatever people want simply because they need it to. “Making” AI do something new in its present stage of development implies a lot of hard work and real research effort, requiring resources that the vast majority of small businesses do not have.

But why?

Because AI technology is currently a “dark art” in the sense that it requires highly specialized talent. There are a very limited number of people with the requisite high-level skills to actually get their hands dirty and build an AI application to drive business value (we try to interview these experts every week on our AI podcast, extracting and translating their insights for business leaders). It is hard science and takes a lot of time – more time than most enterprises can bear, never mind small businesses.

Additionally, building AI applications is not a solo endeavor, it commonly involves team of experts within the business – many of whom must be trained in the novel science of AI. Putting such a team together is challenging even for the largest enterprises in the world.

Data Scientist Salaries

Data scientist salaries aren’t cheap – Screen shot from Glassdoor.com

Since the talent is rare, they are much in demand with companies doing big things with AI/ML such as Facebook, Google, and Amazon. Consequently, AI talent can demands massive compensation, for which most small businesses simply do not have the budget.

Additionally, AI technology requires dat governance and infrastructure vastly more complex than that which most small businesses can afford. These are just two of the basic requirements to leverage AI technology effectively, and for small businesses they are daunting.

Altair 8800

A photo of the Altair 8800 from 1975 – Source: Wikipedia

You can think of modern AI now as the personal computer was in the early 70s. While many small business owners may have been familiar with the term “computer” in the 70s or early 80s, the technology was complex and involved.

These first PCs worked and looked nothing like the computers of today, and only a very few people could program these archaic machines to “do” anything useful at all.

Then, as in now for AI, personal computers were not necessary to run a business. If you had a small business in 1980, such as a bookstore, a couple of restaurants, or a lumber delivery business, you would not need to buy computers and allocate a ton of your time to getting people to learn how to do programming on these new fancy personal computers.

Computer adoption for most small businesses in 1980 was not just unnecessary, it was probably an irresponsible use of time and resources.

The same thing is true for AI in small businesses in 2018. It is not generally necessary to reach a small business’s objectives. Most people in the 1980s did not feel the need to rush out and get computers for their small business, but (due in large part to the hype and excitement around the technology) many small business owners today feel that they should be finding ways to use AI immediately.

Of course, some people in the 1980s did correctly assess that computers are a good investment because they saw how it could help their business in the future. The same exception applies for small business owners today who can actually leverage AI technology to improve business functions and profitability now.

In other words, nobody should be thinking about rushing into AI without a clear business objective. 99.9% of small businesses don’t actually need AI right now to become profitable, and the technology requires time and skills that make it hard to use. As AI becomes more accessible (cheaper, with better defined business uses, and with less requisite computer science knowledge), we should expect to see AI make its way into small business.

Just as PCs and the internet are positively mission-critical for many small businesses today, we should expect AI to be similarly ubiquitous – but not anytime in the next five years.

Leveraging Technologies and Applications that Use AI

That said, small businesses could still leverage technologies and applications that use artificial intelligence. This is by way of companies who can offer AI-integrated easy-to-use technology products. Most small businesses could experience artificial intelligence now by using AI products “off-the-shelf,” so to speak.

Think of it as computer software such as a document editor or customer relationship management. These are readily available and user-friendly. Small businesses have no need to build their own software. They use applications (like Salesforce, MailChimp, Microsoft Word) built by other people who know how the applications work and how to make them efficient.

The same dynamic is likely to take place with AI-related tech. Good examples of how small businesses are leveraging AI today are in programmatic advertising through companies such as Facebook and Google. These technologies are simple to use (and not surprisingly, built by consumer tech firms who are expert in creating easy-to-use technologies).

Facebook TechEmergence

Readers interested in the use of AI at Facebook may be interested in our interview with Facebook’s former Head of Core Machine Learning, Hussein Mahenna

Both Facebook and Google’s advertising platforms use artificial intelligence and machine learning in very robust ways. If you provide Facebook with an email list of 2,000 people for your ad campaign, it will use artificial intelligence to find those users on Facebook, find the commonalities between them in terms of what they like, where they live, what is their gender, what do they do, etc. To create “look-alike campaigns,” Facebook will use an algorithm to target similar people on Facebook to match and cluster these users.

Facebook then uses an algorithm to look at everybody bidding to get exposure to different types of individual users and determine what ads get shown and when. Advertising is spread out across all of those users in real time, and to ads that Facebook believes not only will encourage them to engage and click, but also find useful or enjoyable.

Facebook uses AI and machine learning to provide a better user experience to the benefit of the user as well as for your business. They want to put ads in front of people who are likely to do something with them, and this encourages spending from their advertisers. Google does similar things with its own advertising. You, as a small business, can benefit from these services in a very real way.

Aside from programmatic advertising, you can also leverage AI applications on data security using vendor companies. For example, if your business involves high IT or complicated IT security infrastructure, then you want some serious fraud detecting and security applications to keep your system safe from malicious activity. This is actually a necessity rather than an option for this type of small business.

Small companies are more vulnerable to hackers than large companies are because so many of them do not have adequate checks in place.

One example of an AI-leveraging security application company is Darktrace with its Enterprise Immune System. According to the site, the AI application “builds its own, unique understanding of what ‘normal’ behavior looks like within an enterprise, and can detect emerging threats in real time.” The tool requires no special skills to install it. It is also self-learning, so there is no need for the business owner to fiddle with it at all once it is in place.

Data security is also a serious issue for any business that accepts online payments through credit cards and debit cards. In most cases, merchant account companies have their own security measures in place, and a good number of these have embraced AI technology.  Stripe, for example, is a payment processing company that uses adaptive machine learning to detect fraud. It collects data from all thousands of businesses already running Stripe to improve its ability to detect anomalies and malicious activities for new users. Stripe, too, works out of the box and is easily customizable for any business owner.

Small business owners can leverage AI by using what is currently available from vendor companies out-of-the-box. Many are already using these types of application without knowing anything about AI simply by placing an ad on Facebook or Google, or accepting a payment through Stripe. It is simple, cost-effective, and drives value for the company.

However, you must remember that AI in small business is a means to an end, not an end in itself. Do not be seduced by the hype. Do not use applications just because they use AI. Remember that AI is not magic. It will not solve your business problems or ensure your success by its mere presence.

This applies even when placing ads on Facebook. It may not be a good use of your resources, no matter how small an investment, to do so if you can market your business more effectively offline, such as through a local network or trade show. However, you might be thinking that since Facebook uses AI, it must be a good place to spend your money. That is not the case at all.

AI Enterprise

Readers interested in the “dos and don’ts” of applying AI in larger businesses may benefit from reading our full article on “AI Adoption in the Enterprise

If Facebook is not the best advertising venue for you, then it is not a good place in which to put your ad money, AI or no AI. Use Facebook advertising if it makes sense, don’t if it doesn’t.

Instead of asking where you can use AI in your business, think about what will give you the highest return on investment. Choose software that can help you achieve your business goals. Do not pick software or a company based on what uses AI with the mistaken idea that doing so makes your business cutting edge.

Pick the software that seems to have the best use cases to achieve your business goals and to give a positive return to your company. If that happens to use AI, then so be it. You can leverage AI through existing vendors that are using this for email, security, payment processing or advertising, but no small business should be picking software because it AI is there somewhere.

For one thing, many companies that say they are doing AI are not and those that are really doing AI have products that are too complicated to use for small businesses. For another thing, AI is a buzzword that may not benefit your business. When choosing software or services, think about which software has the features, the prices, and the support that are going help reach your business goals, and drive growth and profitability for your small business.

Prepping for the Future of AI for Small Business

Since AI is eventually going to be the backbone of business just as computers and software are now, companies inevitably want to set themselves up for adoption in the future, and that’s a topic worth considering for small business leaders with significant goals for growth.

There are two very possible circumstances where small companies might get more involved with AI than simply leveraging somebody else’s AI tools that are “off-the-shelf” tools.

In the near future, more and more AI applications will not involve a significant stream of data from small business clients. These applications will likely leverage data from many, many client companies (small and large), and train an AI-based system so that any new client will not need huge reams of data. For example:

  • A marketing automation vendor might determine the best times of day to send certain kinds of promotional emails – or might determine the best way to split-test email subject lines – by working with thousands of eCommerce companies. When a new, smaller eCommerce company begins using the same product, they’re able to begin leveraging this past training data without having to generate it themselves.
  • A security and surveillance vendor might train it’s algorithms to detect suspicious behavior by consuming millions of hours of security camera footage from thousands of clients. When a new client uses this security service, they’re able to leverage all of that training data to detect security threats without providing the data themselves.
  • Etc…
AI adoption poll

Where are business leaders planning to implement AI in the near term? See our business leader AI adoption survey – in partnership with BootstrapLabs

A second circumstance where a small business may have an active role in leveraging data is if it grows to such a degree that they can actually afford to build some of their own AI applications. You may be able  to hire the pricey real-deal hardcore data science and AI people that they would need to build a unique AI application, such as a custom app for your restaurant, or a specialized bit of software to help you with hiring or for an employee feedback system.

Alternatively, AI may become as accessible as many software today, which allows a small business to build its own applications without the need to commit massive resources. AI technology is not yet on par with computer software in terms of accessibility, but it may happen someday.

Realistically, though – as of today – most small businesses cannot actually leverage AI in a significant way today. In most cases, small businesses cannot “do” AI in a way that will provide legitimate near-term ROI. Unless I’m speaking to an AI-specific startup with technical cofounders, I advise most smaller companies to focus on revenue, growth, and data infrastructure – not on “finding a way to use AI” just to fit in.

That said, small business leaders can do something right now to help them prepare for the time when they qualify for either of the two scenarios described above:

Treat your data like a resource.

The need for data is the one thing these two scenarios have in common, and solid data infrastructures will be positively necessary for the AI developments of the future.

AI tools require copious amounts of data to “learn,” so the best way for leaders to prepare for the future with AI is to make sure their data is easily retrievable using compatible formats. You don’t want important business data siloed and distantly separated.

Having a streamlined data infrastructure as part of the DNA of a small business is something that will ultimately serve leaders well.

AI for SMBs – Key Takeaways

Many people will walk away from this article thinking that it did not provide many actionable steps or things they can do to “do” AI.  That is because as a small business there really is not much you can do right now with AI.

Here’s the “TLDR” version of the takeaways from this article:

  • AI is not magic – it will not solve business problems instantaneously
  • AI is still a dark art – it often requires a commitment of massive resources and wizard skills to build applications
  • AI for AI’s sake is a bad idea – choose tools for the value it can deliver to your business, and not because it uses AI in some form
  • Some small businesses can already use some AI applications out of the box – best examples are in marketing, logistics and operations, and fraud detection
  • Data is the key to AI – AI needs data to “learn” and work properly
  • Data is a valuable resource – Small business owners need to store and record data in a consistent manner for eventual use with AI

Again, most small businesses today are not ready to take on AI initiatives. However, that does not mean that you should not keep an eye on AI tools that might affect your industry. Check what the bigger companies in your industry or similar industries is doing with AI, and see if it is working for them. If it seems to be gaining traction, put that on your wish list until such time that you can do the same thing or something even better.  

There will come a time when AI will become as accessible as your favorite computer software, which is when your properly stored data will become invaluable. However, that time has not yet come. In the meantime, the best thing small business leaders can do is to grow their business and take steps to make it profitable. If you have cash in the bank and smart people working for you – you’ll be in a position to adopt and leverage AI as it becomes “small business accessible.”

 

The original transcription and writing for this article (based off of a presentation from Daniel Faggella) was done by  Tini Abadicio, with later editing and additions by Daniel Faggella.

Header image credit: AmTrust Financial

MARKET RESEARCH x INDUSTRY TRENDS

TechEmergence conducts direct interviews and consensus analysis with leading experts in machine learning and artificial intelligence. Stay ahead with of the industry with charts, figures, and insights from our unparalleled network, including executives from Facebook, Google, Baidu, Yahoo!, MIT, Stanford and beyond: