It’s impossible to not have heard of recent advancements in artificial intelligence (AI) in 2023 — or the dramatic pronouncements that have accompanied them. But before you give into the hype, it’s important to know what you’re dealing with and how it can impact your business. To do that, we spoke with David Rombough, Vice President of Customer Experience (CX) Advisory at VXI where he helps businesses leverage advanced technologies like AI to provide excellent customer care. (Full disclosure: David is a Spot On Talent client.)
An Artificial Intelligence Primer For the Uninitiated
You and your company already use artificial intelligence. In fact, you encounter it all day, every day. “AI is everything from tools deployed in the background on phones — looking at photos, photo editing — to tools in my Microsoft Office environment to ChatGPT to how do I take a ChatGPT-like thing to drive search,” says David.
AI comprises a wide range of technologies designed to supplement or exceed human capabilities and powers everything from chatbots to predictive analytics to smart sensors. To help break it down, David recommends putting AI tools into one of four categories:
Data receiving: Visual or audio recognition, for example on your phone or when you speak with your bank.
Transaction processing: Taking what used to be a manual process and automating it through software, such as processing an order through a website or updating you as your order ships. Starting about 10 to 15 years ago, this area has seen significant advancement.
Analytics, Decisioning, Visualization: Using artificial intelligence to process large amounts of data, make predictions or provide visualizations of that data for humans to better understand patterns and predictions. This area of advancement, in the last five to 10 years or so, has been focused on Big Data.
Service Processing: The area where we’re seeing the most recent advancement. It includes language-based AI, inferring emotion and sentiment, translation, generating natural language. Learning through algorithms, deep learning or machine learning. And environmental AI, augmented reality or virtual reality — early examples include the ability to virtually try on clothes on websites, but lately include VR headsets from gaming companies.
So when someone starts using the term “artificial intelligence,” it’s important to understand the specifics of what they’re referring to. For the rest of this article, we’re going to be focusing on large-language model technologies and generative AI in 2023.
What is Generative AI?
One of the big advancements in AI in 2023 is the use of neural networks, natural language processing, deep learning, and large language model based intelligence. These technologies power generative AI. If those terms sound like another language to you, you are not alone. We can help define them:
Neural networks mimic the communication pathways and processes of the human brain. These allow the software to make connections between disparate sets of data, infer contextual information and more.
Natural language processing (NLP) refers to the ability of computers to understand the written or spoken word in the way humans do.
Large language models (LLMs) are a type of neural network that can handle NLP work. Beginning with a huge dataset, LLMs can use deep learning to understand, generate and predict new content, such as text. LLMs are very flexible and can be used to generate images from text, code for programmers, classify and categorize large amounts of information, and more. “[LLMs are] able to access and synthesize data across a variety of different sources and to identify the connections that need to be made across that data,” says David.
Generative AI leverages the above technologies to create text, audio, video, images and more that can arguably pass for human-created work. Some examples include text-based ChatGPT, image creator DALL-E, GitHub’s code-generator Copilot (different from Microsoft’s unreleased Copilot) and more.
7 Things to Know About AI in 2023
- Expect Business Changes Within 12 to 24 months:
While it’s too early to tell exactly what the impact will be, we know change is coming. David says many people are still working to understand how to apply the technology. “We’re still in the lab,” he says. However, the cycles of testing, introducing, and applying AI technologies are getting shorter and shorter. “It just strikes me that our ability to identify the use case, deploy it and move — it happens faster and faster and faster… I think this will start having a massive business impact within the next year or two.”
- Use of AI in 2023 Requires Human Supervision
For one thing, ChatGPT can and will give you incorrect information — with confidence. “I struggle with the stuff that you look at that’s been pushed out there, Chat GPT-4, and the fact that it’s still kicking out errors,” says David. The technical term for this is “artificial hallucination” (we are not making that up). It’s important to note that ChatGPT is just in a beta phase and can and will be wrong, as in the case of the attorney who used ChatGPT for research and ended up citing cases that never existed in court.
- Potential Applications Can Touch All Aspects of Business
Imagine chatbots or audio assistants so powerful, you can’t tell they aren’t human. Real-time coaching for managers, agents and customer-service touch points that not only improves performance but counters problems before they start. “We’re getting to the point where you’re going to be yelling, ‘Representative,’ [when you call support], and what you will get is something that you thought was a representative, and it tricked you,” says David. Going further, imagine AI-generated building plans, AI-optimized go-to-market plans, data normalization and integration without human support. And remember Microsoft Office’s Clippy? A recently announced advancement in Office named Copilot threatens to change how we all do our day-to-day work. It’s nearly impossible to consider all of the places AI could transform our work worlds.
- AI in 2023 Won’t Take Away Jobs; it Will Enhance Them
When it comes to AI in 2023, it can make predictions, infer context and understand sentiment. This could help improve an employee’s output or quality of work many times over. It’s important to communicate the potential value of AI to employees to quell any fears that they may be replaced by the technology. For instance, when it comes to customer service for products, software can reference and make sense of information from internal manuals to Reddit posts and external databases. This helps the customer care staff. “With new tools, I can provide way better data faster to that [customer care] agent in real time,” he says. “We’ve never experienced it. I’m not sure that takes the person’s job away, which is a big fear, but it does fundamentally affect the nature of the work.” While AI will be able to generate text, David says its ability to draft quality content might be limited. “I think what you’ll see over time is it will do the first draft of stuff and then it allows people to clean it up, true it up, make it better.”
- Ethical Implications Run Deep
Of course the use of AI in 2023 comes with a larger social responsibility. Any software or program is only as good as its data — and if the available data is rife with bias against protected groups, which it can be, the program will also be biased. For instance, asking ChatGPT to write descriptions for teacher or mechanic jobs will likely output gendered language. However, this also presents an opportunity to use software to correct against known bias. In fact, at Spot On Talent, we use an AI-based platform called Textio to ensure our job descriptions, communications and more are free of bias. Many states are considering regulatory laws for the technology — New York City has already issued a rule on using automated technology for hiring decisions. On a federal level, standards or regulations are currently being studied by more than one agency.
- If You Don’t Have a Policy Covering AI in 2023, You Should
The FTC stated that current consumer protections also apply to the use of AI in 2023 For the most part, ensuring compliance with privacy protections, equal opportunity laws and more is currently up to companies themselves. Additionally, publicly available tools are not secure when it comes to corporate or customer information. You should consider developing a policy covering the use of AI in 2023. Motivated employees can use the currently available tools to their (and your company’s) benefit, but it’s important to provide guidelines on their use.
- A Rising Tide Lifts All Boats — And Lifts Consumer Expectations
Whether you are ready or not, AI is coming to transform your business. Most companies may not be prepared to (and probably should not) make significant investments in the use of AI right now. However, some organizations have the resources, infrastructure and expertise to dive in immediately. That means smaller companies will have to play catch up to stay competitive and retain customers. Remember when Amazon transformed the type of experience consumers expected — immediate shipping, easy returns, interactions that are customized to their wants and desires? AI promises to do the same thing. Have a customer service issue and need to talk to someone? Gone are the days of automated voice prompts or unhelpful chatbots. Soon, the automated assistant helping you will know more about the reason you’re calling than you do — before you call. While you may not be able to implement this kind of technology now, it should be on your radar.
While a lot of the current talk around the sector of AI is full of hyperbole, it’s important to stop to consider the best ways to prepare for new advances in AI and consider its real-world implications.
