Article updated: September 2023
Over the past year, the search landscape has been changing more quickly and dramatically than ever before. In our Future of Search series, Rosie Hopes has been unpacking the latest developments and what they mean for marketers. Today’s topic: chatbot search.
Unless you’ve been living under a rock for the past year, you will be aware of ChatGPT and its meteoric rise to fame in November 2022. It became the fastest growing consumer application of all time (before Threads beat its record in July 2023) gaining more than 100 million users in the two months after launch. For countless people all around the world, signing up for ChatGPT was their first taste of AI-driven chatbot search.
How Can AI Be Used in Text Search?
AI technology is revolutionising the way we search online. Rather than spending time trawling through endless irrelevant results, AI can help you discover the relevant answers quickly and effortlessly. It can identify keywords and phrases in the data, tailor website rankings to suit your exact requirements and even adjust outcomes based on past searches. Within a few clicks, you’ll be able to find precisely what you’re after.
What is ChatGPT?
ChatGPT is a particular kind of natural language processing (NLP) that allows conversations with artificial intelligence-driven bots. By assessing the context of a discussion, ChatGPT can answer questions, make recommendations and even complete tasks such as customer service or sales. This groundbreaking technology is quickly gaining traction in chatbot design because of its highly engaging and personalised user experience. It can also be used for larger data collection projects for analytics and research purposes.
And, considering the topic of this section, it may not come as a massive surprise to you that… the last two paragraphs were written by AI. I made some minor tweaks – I have an aversion to both unnecessary exclamation marks and the word ‘moreover’ – but I pretty much took it verbatim from the suggestions given by Jasper, an AI content creation tool that we have been using as an agency for a few years now.
The Battle Between Google and Bing
AI chatbot technology has been around for years in various applications – usually gated by paywalls – but we’re at an important breakthrough moment now as both Google and Bing are incorporating it into their free search offering.
The New Bing
Many commentators are wondering whether Microsoft’s acquisition of OpenAI, the company that developed ChatGPT, is their most significant challenge to Google’s dominance in the search arena to date. Satya Nadella, Chairman and CEO of Microsoft, said the following at a press conference on 7 February 2023:
“AI will fundamentally change every software category, starting with the largest category of all – search.”
As of March 2023, Bing rolled out access to its new “AI-powered co-pilot for the web”. When you type in a search query, Bing will generate a traditional SERP with featured snippets, as well as offering an AI-generated response that you can use as a starting point for a chatbot conversation.
In my opinion, a welcome addition to the chatbot search result is a list of citations showing the sources for the information used. This is a feature missing from other chatbots and earlier iterations.
Not to be outdone, Google announced the launch of its own AI chatbot, Bard, the day before Microsoft. However, Google stock prices were hit when an incorrect response from Bard caused consumer trust in the technology to wobble. Bard is now available in more than 230 countries in over 40 languages at https://bard.google.com/. Unlike Microsoft, Google hasn’t yet integrated its chatbot search function into the regular Google search engine. It makes it very clear from the outset that Bard is an experiment that may not provide accurate results.
Interesting fact: when I first wrote this article in February 2023 and asked Jasper “what is Google doing to compete with AI chatbots?”, it told me about Google Duplex, which is its voice-based AI offering. It was announced that Duplex is being retired from the web in December 2022. So Jasper missed both the Google Bard announcement and the retirement of Duplex and as such gave me an answer that was completely out of date in a fast-moving industry.
This exposes one of the main hurdles of the current crop of AI tools: keeping its index up to date. Bing Chat had this to say on the subject when I asked the chatbot “How up to date is the Bing chat database?”:
The Bing chat database is updated periodically based on the latest data and models available. According to the Microsoft CEO of Bing Search, Mikhail Parakhin, Bing Chat’s models take several months to train, so users can expect the models to be updated maybe three times a year or so. The models are updated outside of some small RLHF tuning runs. The last update for Bing Chat was on June 2, 2023, when the number of chat turns and the image creator feature were improved.
Google Bard said the following:
The Bard database is continuously updated with new information from the web. This means that the information I have access to is generally up to date, but there may be some lag time for some topics. For example, if you ask me about a recent news event, I may not have the most up-to-date information if it happened within the last few hours.
This will be an important development to watch in the coming months and years. While Bing may have taken the lead with integrating chatbot search into its search engine, will Google’s more regular database updates make its answers more accurate, and therefore more sought after?
How Do These Developments Affect Marketers?
I could write an entire essay on AI chatbots alone – their benefits, weaknesses, developmental issues and the ethical quandaries they present, especially for education. But I want to keep my focus on how they may affect search and the people who harness search engine marketing for a living.
I think it’s very important to reiterate how new this technology is in the public search sphere. We can’t confidently say how it will affect the average person’s search habits, or what marketers can do about it. AI chatbots are not perfect fonts of absolute fact and wisdom, and the technology has a long way to go in terms of developing, learning and growing. I think that Sam Altman, the CEO of OpenAI, says it best:
I believe that there is one actionable insight that search marketers can take away at this point in time:
Chatbots can only analyse what they find in their index. And their index is *the internet*.
Without web content, chatbots are useless. They only work if brands, institutions, publishers and individuals are continuously adding new, fresh, accurate information to websites. The relationship has to be symbiotic. (If you’re interested in my personal take on why we still need human writers, take a look at this blog post.)
It can be easy to use new AI tools to make content production quicker and cheaper, but I would urge brands not to rely wholly on using AI-generated content on their own websites. Instead, think about how to provide your audience with a fresh take, a tailored response or a new insight. Rather than churning out four generic AI-made blog posts a month, concentrate on one human-made blog post that offers something new to the user – and the internet.
And if users decide that they prefer finding information through chatbots and our websites stop getting organic traffic? We’ll cross that bridge when we come to it because human beings are excellent problem solvers.
There’s still a lot more to know about the developing search landscape. Download our Future of Search whitepaper to learn about the most recent innovations in social, visual and voice search. If you still have questions, we’re here to help – get in touch for a chat.