The Five Most Key Takeaways from This Blog
- OpenAI’s ChatGPT may be the most popular chatbot in the world, but for reasons that may not be relevant to business owners in the way IBM Watson is. The wide range of tasks you can set to ChatGPT is impressive, sure. Who does not find an A.I.-generated example of a Nordic warrior-elf offering a haiku greeting to a friendly dragon (see below for an example)? But business owners need narrower conversations for chatbots, as opposed to wildly and variedly creative.
- IBM Watson (currently in a rebranding phrase as “IBM watsonx”, fashionably lowercase) is attractive to business owners because it has a longer history of real-world, real-time applications in the businesses, including at the global-enterprise level.
- One significant use for chatbots is in customer relations and customer outreach. This can take place on a website, over social media, email, or other text-based computer-mediated-communications.
- IBM Watson/watsonx’s customizability is specifically made for business contexts. That encompasses both text and voice assistant integrations.
- Major companies like Humana and Citibank have seen success with IBM’s watsonx A.I. assistant.
The Race to the Perfect Chatbot
ChatGPT had a big public moment in November 2022 when its developer, OpenAI, made it open to just about anyone and everyone with an Internet connection.
A decade-and-a-year earlier, IBM Watson made a similar big-splash public debut when it competed on Jeopardy—and won!
Both are examples of conversational computing platforms performing at a high level.
Much of the work here involves Natural Language Processing (NLP) and information retrieval. A question is a prompt that sets the chatbot’s algorithm into motion: Analyze the data, find significant patterns, predict what the output should look like, and generate that output.
Sounds simple, and it really kind of is, at least in the abstract.
ChatGPT Has Made Its Mark
So, why has ChatGPT captured the public’s imagination for a sustained period of time? More than a year, which is an eternity-and-a-half in Internet Time, has passed since ChatGPT’s debut, yet its name is still on everyone’s lips.
Yet, IBM Watson’s Jeopardy was yet another example, much like the famed Turing Test, that gave people a profound glimpse at the power of AI, yet people sort of…moved on, waiting for another example of AI to blow their minds at a later time.
Now, ChatGPT has made a lasting impression, one that has shown the world just what AI is capable of.
Perhaps the strongest explanation for this difference in sustained-impression-making is that ChatGPT is/was interactive so that users could get firsthand experience of interacting with AI.
Meanwhile, IBM Watson’s Jeopardy run was more spectacle than interaction.
However, that spectacle made a deep impression on one group of people: business owners across a variety of industries.
How Many Business Owners Use IBM Watson?
Major companies like Humana, which has one of the largest healthcare companies in the world, have seen success with IBM’s A.I. offerings.
Humana was looking to lower the amount of calls that were being sent to human agents just to confirm pre-care coverage. These types of calls are, as you can imagine, not exactly the kind of involved customer-service conversation that a company’s human worker necessarily needs to be on.
The solution was to integrate Humana’s Voice Agent with IBM watsonx. That led to handling more than 7,000 calls on a daily basis.
Citibank, which is a multinational banking company, has been using IBM watsonx to help with internal audit processes. An example here includes scanning through literally hundreds of pages to identify relevant content in a document.
Worth noting as well is that “relevant content” is something that businesses can indeed train A.I. to spot. What may be “relevant” to one company may be irrelevant to another, hence the need to tailor an A.I. to a specific company’s needs. That is why so many businesses find IBM watsonx’s customizability a great feature.
What Makes Watson Quality and Reliable?
ChatGPT is a conversational computing platform that can do quite a bit, but is not exactly masterful at anything.
IBM Watson, on the other hand, has been trained to hit very specific conversational targets, such as having a successful conversation with a bank’s customer seeking information about a loan.
When ChatGPT makes a flub in a poem, it is no big deal. It is charming, even. ChatGPT is not perfect, but for the purpose of having a wow-factor technology, it is good enough.
By the way, we did mention in the Key Takeaways section that we would include the haiku greeting by the Nordic warrior-elf. Here is what ChatGPT generates:
Flames greet forest breeze,
Ancient friends in twilight’s glow,
Strength and peace unite.
Cute enough, sure, but given the track record of making errors, do you really want ChatGPT speaking with your customer base? Imagine if ChatGPT gave a bank’s customer inaccurate information about possible interest rates for a loan.
With IBM Watson, which has been fine-tuned over the past decade, you can rest easy knowing customer interactions will go more smoothly.
