The question of trust, which I wrote about in a previous blog is becoming more important by the day.
Recent analysis from two companies, Unicepta and Muck Rack, which landed in our inbox shows the extent to which external sources and narratives are driving people’s views of companies.
Unicepta found that in over 90% of its analyses1, the way in which companies are represented in AI systems is largely driven by external sources and narratives.
Similarly, Muck Rack analysed more than 1 million links from AI responses2 and found that non-paid media accounts for about 94% of links cited; the majority of which come from journalistic sources. Earned media (non-paid journalistic coverage) forms about 25% of all citations.
So naturally I wanted to ask my new best friend, ChatGPT, what it thought about all this. I asked how many of its own citations are driven by journalistic sources, and it said: “journalistic sources typically account for somewhere between ~20% and ~47% of AI citations, depending on the study, query type and methodology.”
You could surmise that this is because journalist-generated copy is trusted. Journalists – and publications’ sub-editors – are a reliable third-party voice when it comes to facts. Facts have been checked, sources verified. The element of trust is there.
This is important for any brand because people are increasingly starting to switch their search activity from traditional search engines towards AI models3 – a trend which is set to gather speed exponentially if my own search use is anything to go by. Generative Engine Optimisation (GEO) is usurping SEO when it comes to search.
Clearly this shift will have implications for companies who want to raise their visibility. Credible media coverage will be more important than ever if you want your company to be considered or even discovered by audiences.
Quill PR has been helping clients raise their profile via earned coverage for over a quarter of a century now, and we understand how positive media coverage can help businesses meet their objectives. The importance of this type of exposure looks set to grow exponentially over the coming months.
Chat GPT then helpfully offered me a ready-made quote I could use in my blog. Thanks friend!
“Roughly a quarter of AI-generated citations now come from journalistic sources, rising to nearly half for time-sensitive queries such as markets and regulation. For UK financial services firms, that reinforces a familiar truth: authoritative, earned media coverage isn’t just shaping reputation — it’s increasingly shaping how AI systems interpret credibility and surface information to investors.”
Emma Murphy is a Director at Quill PR
1. source UNICEPTA: GEO – Generative Engine Optimization for AI Search Visibility
2. source What Is AI Reading? December 2025 [Final]
3. source 37% of consumers start searches with AI instead of Google: Study
The march of AI or artificial intelligence is truly upon us, highlighted by the fact that as I type, my keyboard is predicting what I am going to type next, and is often correct. Gosh, so predictable!
At an extremely interesting client presentation recently, fund managers from the Sanlam Artificial Intelligence fund explained to us that “Things that people called ‘pipe dreams’ when the fund was launched in 2017 are now happening. The changes have been utterly extraordinary. What was impossible is now possible.”
A fascinating timeline showed how algorithms have achieved superhuman levels in chess and other games such as Go in a very short space of time, and how AI platforms were teaching themsleves to walk, create their own languages and understand human vocabulary at an increasingly exponential speed.
There’s no doubt that changes brought by AI within healthcare, agriculture, transport (to mention a very few) are literally saving lives and may help to save the planet.
However, the very night after this event, I was reminded about the limits of AI in the arena of customer service. One area where chatbots and robots haven’t quite nailed the human touch yet.
My teenage daughter and her friend had arrived at a hotel in Portugal late at night, to be told that their room booking had been cancelled by the online booking service they’d used months earlier. This was human error (or greed) on the bookings service part, no robots at fault there. However, her booking was confirmed so a bit of a disagreement ensued but as the hotel was now full, she needed to contact them as they now had nowhere to stay. So then she (in Portugal) and I (in London) tried to communicate with the bookings company to get the issue sorted. We both experienced the same fate…
The phonecall was answered clearly by a robot – who requested the confirmation number. That was easily done. There then ensued some fake typing noise, I suppose to suggest that someone was actually typing into a computer to check something out, before a robotic voice informed that they had been waiting too long for our information and would have to end the phone call. At gone midnight – and having gone through this process twice I was definitely not impressed. The situation was not resolved until the following day – and took human intervention via twitter direct messaging to sort out.
While AI is clearly the future, companies should beware that they are not jeopardising their hard-won reputations for short-term cost savings. The message is that the nuances of real life problems often need to be resolved by humans; and human customers are not happy when they have to battle with companies’ attempts to deflect issues to our robot brethren, before they are quite ready.
Photo by Alex Knight on Unsplash