We are currently witnessing the dawn of large language models (LLMs), such as ChatGPT. These are changing the way we work and the way we learn – particularly the way we search for information. There has been a huge reaction from education leaders, worrying about such tools being used to help students cheat their way through their studies, or fearing that they will be fed incorrect information.
On the other hand, in the workplace, GenZ employees have bought into the AI hype. They are using the technology to help them with various work tasks, but have a huge fear of managers finding out. This is due to lack of company regulation around whether they should or shouldn’t be using these tools to support their work.
The real conversation here though is, how useful is ChatGPT and other similar tools when it comes to research? With over 80% of the search market share, Google is the household favourite, but even Google has its limitations. Google is set up to search by keywords, but not to dive into granular and complex questions. For example, if I use Google and search ‘AI’, the results come back with a multitude of news items, various descriptions of AI and a range of company articles using the term ‘AI’.
This is where tools like Chat GPT come in. Using an LLM, I have the ability to ask a question such as ‘Can you describe what AI is’, and it comes back with a detailed description of AI and its use cases. This is information that can be pulled into any written work without having to use a single brain cell. This type of language model has the ability to understand and respond to natural language and provide answers that are both informative and entertaining, generating a variety of responses to each user’s questions.
However, the major limitation of ChatGPT is that the data only runs up to 2021, so for many trying to use this tool, the information will be far too out of date to create current and reliable content. This is a major point for those working in tech comms, as the speed of innovation is so fast that information quickly becomes outdated.
Aside from this limitation, there have also been concerns around the ethical implications, including privacy, bias in training data and lack of human interaction. More commonly used search engines don’t have these same problems, and therefore are more reliable to use for research. Using a manual search engine relies on people to manually gather and organise their own data and information, based on the latest information available. On the other hand, an AI search engine relies on computers and algorithms and their pre-trained and installed data to produce results. This is one of the key differences when using either for searching.
However, a search tool is only as good as the data it provides. Google provides results to our keyword searches based on the algorithm it uses to deem information credible. ChatGPT hasn’t yet been transparent about its sources, which again makes using it for research difficult.
Looking at this from a comms perspective (as we’re comms people after all) these changes will be significant to our output. Firstly, we’re constantly researching to ensure we are knowledgeable for our clients. But secondly, and importantly, a lot of what we do influences Google results. An amazing article about our client in a national newspaper like the Financial Times, will feature at the top of search results and will have an impact on that company’s reputation. In B2B, the sales process often starts with Google! But as LLMs continue to develop, what will it mean for a company’s reputation and how they feature in LLM results?
There is no doubt that LLMs will continue to have a huge impact on the way we search, work, and learn. We’re at an important juncture, where not only the likes of Google will look to make significant changes to its platform, but we’ll also see a huge range of new players enter and compete in the ‘AI race’. It’s not too dissimilar to when we witnessed the disappearance of Nokia, Motorola and Blackberry as Apple and Microsoft became the dominant players in the mobile phone evolution. I think we’ll see something very similar happen here!
The evidence is right there. Virtual assistants are here to stay.
The last time you spoke to Alexa, Cortana, Siri or Google, it may have felt a little strange, but tech companies are pouring money into developing voice technology and placing their bets on the major impact it’ll have on our daily lives. Innovation in virtual assistants is all about making them as human as possible – Amazon’s Alexa recently announced how it rewards children who say ‘please’, whilst Match launched chatbot Lorna on Google Home, making ‘Lorna’ a source for dating advice, as well as your friends and family. It may seem a little surreal today, but the more intuitive and easy-to-use the technology becomes, the more likely we are to fully embrace it.
As voice technology develops, marketers must prepare themselves and start to think about how to optimise PR and marketing content for voice search. Wondering why exactly?
Much like search marketing changed the value of PR, so will voice. As a reader of Firefly blogs, you’ll know that for some time now we have put a large emphasis on optimising PR for search, given the changes in search algorithms from both Google and Bing. Essentially, PR brings SEO value when you target the right media sites with the right content, and get that sought-after link to your website. It’s not just about awareness and share of voice, improving SEO is a legitimate PR measure too.
Of course, voice search is a completely different beast. For starters, the way you type something into search engines and the way you speak to a virtual assistant are completely different. We’ve adapted to speak to Google in a way that we get the information we want as quickly as possible – ‘burritos Farringdon’, for example, if you’re on the hunt for some Mexican for lunch. Whereas virtual assistants we expect more, and we don’t naturally speak like that, we’re more likely to say, ‘where can I get a burrito near here?’
Then there is the number of devices and voice technology applications – it’s not just a case of monitoring search engines’ changes in algorithms. Here is an extract from our latest whitepaper on PR in the age of voice search which highlights this complexity:
When Alexa is asked who the CEO of Facebook is, she gets the answer correct – Mark Zuckerberg, of course. When she is asked where she got that information from, she says that ‘that’s tough to explain’.
Different devices use different applications to source information. With Siri for instance, the same Facebook CEO question pulls up the answer on ‘Knowledge’, through Cortana the answer is pulled up on Bing and with Google Home, unsurprisingly it comes through Google search.
A location-based question like ‘where is the nearest Italian restaurant?’ will use a maps application to give you various options.
The market is also in a state of growth and development, and it is not clear yet which virtual assistant will maintain dominance. In 2017, Amazon and Microsoft announced a partnership in order to strengthen Cortana and Alexa by combining both virtual assistants. Essentially, within Alexa you’ll be able to open Cortana and vice versa. The idea is that people can use the functionality in both assistants like sending an email in Cortana or using an Alexa skill to order a pizza.
With some of the virtual assistants like Siri and Cortana, your search marketing efforts will already put you in a strong position. Asking a question such as ‘Who are the top entrepreneurs in the UK?’, Siri and Cortana will refer you to the top web results. However, with devices like Amazon Echo and Google Home being ‘screenless’, there is added pressure to be THE answer that is read out. And you thought being on page one of Google was tough!
Though the market is still in a state of flux, key players are emerging and there are ways you can prepare today – all included in our whitepaper, of course.
As search marketing evolves, the most important thing is to be ‘found’ by people genuinely looking for you. There are so many ‘secret’ quick wins with voice search, much like online search, but you have to build and maintain your profile so that when the opportunity arises, you reach your customers.
Want to know more? Download our whitepaper on PR in the age of voice search here.
AnswerThePublic isn’t a new tool, but it’s a goodie if you’ve not come across it before. The site collates the auto suggested results provided by Google and Bing, for a word or phrase. It’ll search keywords whether you want to look for the name of a product or service, the name of your brand, or your competitors. The clever people at AnswerThePublic also created a way to visualise the data, grouping questions by who, what, where, when, why and how.
In the example on Facebook, you’ll see from a quick scan that Facebook has some work to do on how it explains its privacy settings. You’ll also notice topical questions pop up, like in this example ‘Which Facebook friends like Trump?’, so it’s always worth refreshing this insight on a regular basis.
This search insight is valuable for PR and marketers in many ways – whether you want some prompts in a creative brainstorm or you want to understand how your competitor is being searched for. Also, in the safety of their browsers, people are more likely to express their true feelings and preferences, and you may uncover insights you won’t get from surveys. Often in surveys you get social desirability bias – meaning that people over-stress good behaviour and under-stress bad behaviour. You don’t get that when you’re analysing search patterns.
Search insight is also especially interesting as PR and SEO become more intrinsically linked. Google is constantly tweaking its algorithm to point people to the right answers, so it’s particularly useful to know which are the popular questions associated with your brand and products/services.
So, next time you’re searching Google for ‘is there a tool that collates auto-suggested search results?’ you already have your answer!
All of the major search engines are now very much social: Bing is happily in bed with Facebook, while Google, after a lot of flirting with Twitter, is risking leading its own social initiative with Google+. We at The Search Agency know the search engines want to buy into the social space; but where does the integration opportunity lie for brands?
What is clear already from the existing venture between Facebook and Bing is that the ‘search’ results within Facebook are now very much spilt between actual Facebook results and Bing results. And users are aware of this. People are ‘checking in’ all over Facebook and as they do this, not only are ads thrown into the mix, but search results are becoming increasingly targeted. It’s not terribly sophisticated, but users are interacting so we, like most, are beginning to conclude that Facebook and Bing are stepping in the right direction.
And now there is Google+ and the huge expectation that it will substantially boost Google’s ability to aggregate social activity and relevant search results. Google does search very well. Heck, Google does video, maps, mail and news all pretty well. To date, Google has not done the social thing terribly well – with the exception of YouTube, of course. And this is where the lines become a little blurred as YouTube is the second largest global search engine. Will Google+ prove Google has learnt a lesson from Buzz, Wave and other forays into social? Only time will tell if users will engage, but maybe Google should be swotting-up on more of what it does well and building this into their ‘social engine’.
Social campaigns will still lean towards Facebook, Twitter and YouTube in the latter half of 2011, but surely Google has learnt a thing or two by now. Can the number one search engine become a real social engine? We will just have to see!
Receive thought pieces from our leadership team, views on the news, tool of the month and light relief for comms folk