Artificial intelligence (AI) used to feature heavily in science fiction movies but rarely in business IT. These days, it's a commercial fact. The evolution of the cloud, along with the creation of new microprocessors that specialise in training AI, has driven AI into lots of businesses. It began with enterprises large enough to afford it, but now small companies are getting in on it too.
AI captured the attention of large and small businesses alike because it works differently to traditional software. Instead of following distinct predefined rules, it uses complex statistical models that emulate the way the human brain works. This enables it to deal with complex data like images and text in ways that traditional computer software can't.
These capabilities enable AI to automate things that computers couldn't handle before. Small, repetitive, but cognitively complex manual tasks, like reading and analysing simple documents or conducting visual product quality checks, were things that only people could do. That's no longer true.
The first piece of advice with AI is not to try rolling your own. Building AI involves training algorithms using massive volumes of data. That training process takes a lot of computing power, which is expensive to buy on-site. Even if you train an AI algorithm in the cloud, you need data science skills to continually refine it. Most small businesses can't find those skills, or afford them.
Luckily, software vendors have been busy training their own AI and building it into their products. This puts AI-powered algorithms in the hands of any small business using common off-the-shelf software or cloud-based applications. It can help you in a variety of use cases.
Sales teams face difficult questions when trying to target customers. Which leads should they prioritise first? How qualified are those leads? What's the best way to contact those customers? An AI-powered customer relationship management (CRM) or marketing tool can help with tasks like these. It can analyse social media posts to score how customers feel about your product, finding the customers most likely to buy from you right now. It can score the likelihood that a customer will open an email based on past communications activity, and select the best email text to move a deal forward.
One of AI's clearest applications is in natural language processing (NLP), which enables it to recognise human language rather than relying on complex data query languages. Companies like IBM and Microsoft offer cloud-based services that allow small businesses to build NLP-enabled chatbots.
These online messaging bots can field customer questions and handle basic tasks like product questions or account queries. They enable your business to operate 24/7 online, even if you only operate 9-5 on weekdays in the physical world. They require next-to-no coding thanks to clear tutorials and low-code tools.
A chatbot doesn't have to be just a customer-facing thing. It's one of several tools in an HR manager's arsenal. A chatbot can help companies to answer routine questions for employees, and can even front a self-service portal to deal with staff tasks like finding appropriate holiday times or handling benefits issues. AI-powered learning platforms can also help companies to plan training paths for individual employees.
AI has great potential for recruitment. Its ability to handle repetitive tasks too difficult for traditional programs makes it useful for sifting through resumes and find only those that matched key characteristics, selecting candidates more quickly and accurately.
AI can help small businesses in many other ways, including financial planning (analysing numbers and recommending financial decisions, such as whether to make a capital purchase or pay down debt), and fraud prevention (watching for suspicious transactions). Small businesses can begin with small, discrete investments in this technology to help automate clear tasks that complement their human skills.
You don't have to tackle grandiose projects with this exciting new technology. In fact, now that it's embedded in so much cloud-based software, some of its outcomes are so subtle that you might not even realise you're using it at all.