Artificial intelligence (AI)

If you run a business, you’ve almost certainly wondered if you should be trying to integrate artificial intelligence (AI) into one or more processes. It seems like from self-driving cars to chatbots to text generation, AI is starting to handle many tasks at least as well as humans can—and faster.

Before you leap into it, make sure you understand a few things about AI, including what it can do right now, what the best use cases are for businesses, and the pros and cons of introducing it.

Why is AI important?

According to Samuel Easto-Lefebvre, a Senior Business Advisor with BDC Advisory Services, AI is important because of its ability to work through massive amounts of data quickly to produce information that can benefit people. An easy example is medical imaging: AI algorithms can analyze vast datasets of medical images—such as those taken by X-rays, MRIs and CT scans—with great precision to identify patterns and abnormalities that human radiologists might miss. This can lead to more accurate diagnoses for patients.

AI is being put to work in sectors from health care to finance to transportation, agriculture, energy and more. For businesses, it can save time and resources by automating repetitive tasks, producing useful analyses to support decision-making, and more.

How does AI work?

You likely already use a few different AI technologies in your day-to-day life, says Easto-Lefebvre. For example, Google Maps is a form of AI, as are smart home devices (think Siri or Alexa). Personalized social media content is the result of AI algorithms, as are the movie and music recommendations you receive. What they all have in common is a reliance on data analysis, machine learning and sophisticated algorithms to learn your preferences and deliver more personalized experiences.

Virtual assistants, health and fitness trackers, and language learning apps are further examples.

What are the main types of AI?

AI specialists tend to speak in terms of three broad types of AI:

• Weak/narrow AI, which is designed for specific tasks and lacks general intelligence (that is, human cognitive abilities)
• Strong/general AI, which possesses human-level intelligence and can perform as wide a range of tasks as humans can
• Superintelligence, an advanced form of AI that surpasses human intelligence and capabilities, potentially posing unprecedented challenges and opportunities for society

You might be surprised to learn that despite their obvious power, most of the AI technologies you’ve encountered to date are forms of so-called “weak/narrow” AI: they are designed to accomplish or automate a specific task, but don’t have the ability to reason on their own.

When you think about sci-fi movies that involve robots gone rogue, or computer systems that can formulate their own thoughts, that’s “strong/general” AI or even superintelligence, says Easto-Lefebvre—and we don’t have those yet. They’re still in the research phases.

Three “weak” AI technologies that are in common use today include:

1. Generative AI

Generative AI can create new, unique outputs—like images and human language—from provided inputs. Examples include conversational AI tools (like ChatGPT), chatbots and some virtual assistants. Natural language processing (also called NLP) is an AI technology that plays an important role in supporting some generative AI capabilities, enabling computers to interpret and generate human language so they can communicate with humans in ways that sound natural and are meaningful and contextually relevant.

Generative AI can also create images from text descriptions, including photo-realistic imagery and paintings.

2. Machine learning

Machine learning refers to the development of algorithms and models that enable computers to learn from—and make predictions or decisions based on—data. The models are programmed to perform specific tasks, and once trained, no longer need explicit instruction to make decisions. Machines improve their performance of a given task through the experiences they gain from data.

Deep learning is a subset of machine learning that focuses on deep neural networks (networks with at least two layers of complexity that process data using sophisticated math modelling). It is meant for more complex situations like image and speech recognition.

3. Computer vision

Computer vision is a field of AI that enables computers to interpret and understand visual information from the world, such as images and videos. It involves the development of algorithms and models that allow machines to process, analyze and make sense of visual data so they can recognize objects and faces, classify images, and even navigate autonomously by "seeing" and interpreting their surroundings.

For example, computer vision can help diagnose cancer from imaging or be used in agriculture to tell whether a fruit or vegetable is ready to harvest.

What benefits does AI bring to the workplace?

But while generative AI, machine learning and computer vision are among the most recognized AI technologies today, they are not the only ones in use. Others include robotics, recommendation systems, speech recognition and synthesis, and autonomous systems.

Some consumer products use multiple technologies together. For example, self-driving cars use machine learning, computer vision, and other components like sensor fusion and control systems.

Alone or combined, AI technologies have three key potential benefits to offer small to medium-sized businesses, says Easto-Lefebvre. They can help to:

• Cut costs by automating routine processes and repetitive tasks
• Increase revenues by using predictive analytics to produce information, such as which customers are more likely to buy products or what markets a business should pursue
• Add value by enabling existing products to perform better

“AI can cut costs by automating repetitive tasks such as scanning and processing invoices, thereby reducing manual labour and increasing efficiency,” says Easto-Lefebvre. This could potentially free up employees to do more interesting work, in turn helping you to attract and retain talent.

“Likewise, if you manufacture a product that needs servicing on a regular basis, you could add value by using AI to alert the user when maintenance is needed.”

How is AI used by businesses today?

Samuel Easto-Lefebvre says some ways to incorporate AI into a business include:

• Chatbots for customer support
• Sales tools to suggest cross-selling or upselling opportunities based on customer behaviour analysis
• Data analytics to identify market trends and inform decision-making and forecasting
• Inventory optimization to predict demand, reduce overstock and make sure the right products are available when you need them
• Recommender systems that enable users to make better decisions
• Personalized emails for communicating quickly with certain individuals or groups
• Employee training and development
• Cybersecurity
• Social media management and website optimization

Supply chains and seasonality offer a good example of how AI inventory optimization can work.

“Consider a retail company that sells seasonal products,” says Samuel Easto-Lefebvre. “By using AI, it may be better able to forecast which products to order first—and when—in order to have those on the shelves exactly when customers are likely to start walking in to buy them.”

Although there are a wide range of uses for AI, he adds, most small and medium-sized businesses will get the best results using it to automate repetitive tasks.

What risks are associated with AI?

But Easto-Lefebvre adds a note of caution: start small.

“When you look at use cases, not only should you look at the cost and return on investment (ROI), but you should also start with something digestible—something you can fail at without having too much impact,” he says. “Something that is also scalable, so that as you learn from your mistakes, you can continue to integrate the AI and discover new ways to use it.”

You’ll also want to keep an eye out for several different risks when using pre-built AI tools (such as ChatGPT), says Samuel Easto-Lefebvre, including:

• Data privacy. For example, an employee who types proprietary information into ChatGPT could be breaching your company’s privacy. The data could be fed into AI systems later.
• Copyright issues. Generative AI technologies sometimes provide users with copyrighted information in their responses—without accreditation or attribution.
• Bias and inaccuracy. Because you probably don’t know what data a given AI technology was trained on, you shouldn’t fully trust its outputs. “If the data used were not diverse enough, the answers you get may be skewed,” says Easto-Lefebvre.
• Expecting too much. For example, if you use AI to generate your HR policy, and don’t review it before putting it to use, that’s risky. “You still need an HR specialist to verify the output, because you can’t trust it 100%.”

Beyond these risks associated with pre-built AI tools, you also need to be aware of potential implementation challenges, says Easto-Lefebvre—starting with your own corporate culture. AI technologies do their best work when a business already has a culture of “continuous learning”—that is, where employees are encouraged to educate themselves about data, continuously look for AI opportunities, and be aware of the pros and cons.

Your business may also be well-positioned to benefit from AI if it has a history of adapting well to change, if employees tend to be enthusiastic about new ideas and willing to take reasonable risks, if you have effective systems in place to communicate with employees about change, and if there is a high level of trust in your organization.

Thoughtful change management during implementation is also critical, says Samuel Easto-Lefebvre. You can implement all the AI you want, but if employees don’t trust it or are worried that it could replace them, they may refuse to use it.

Finally, be aware that many AI systems, especially custom systems built internally, need to be maintained and constantly updated, requiring an ongoing investment of time and funds.

What should you do with all this information?

If you’re tempted to explore AI, but are still not clear on where to start or how it could help your business, Samuel Easto-Lefebvre says there are two things you can do. One is research: look online and ask around. Find out what technologies businesses like yours are using successfully. Read websites and technology and business magazines to get ideas about AI use cases you would like to integrate into your business.

“Most smaller firms would be best advised to start with ‘out-of-the-box’ AI solutions, such as those embedded in the platforms or software packages they already use,” says Samuel Easto-Lefebvre.

The other approach is to seek professional help. There are AI consultants who can size up your company and its goals and help you put together a business case for the use of certain AI technologies.

“If it's a strong use case, a consultant can also estimate your ROI and even go all the way through to implementing the AI using your company's data or third-party data,” says Easto-Lefebvre. “Some consultants can even maintain the AI for you on an ongoing basis.”

But above all, he emphasizes, take a step back and take a hard look before you spend a lot of money incorporating AI.

“Really take the time to identify the best use case for AI and to understand the cost versus return, because it does come with costs,” says Easto-Lefebvre. “And don’t try to use it just for the sake of it. AI can be a very powerful tool—but it isn’t the solution for every problem a business faces.”

Next step:

Discover how to build a digital strategy with BDC’s free guide Making Tech Work for Your Business