Artificial Intelligence

OpenAI upgrades ChatGPT search with shopping features

Users can now search – and buy – products through Chat GPT, Open AI has announced. 

ChatGPT will now offer product recommendations, images, reviews and include direct links for users to buy the products. 

This doesn’t include paid-for results but instead organic search results with plans down the line for AI to feed relevant results to users based on prior searches. 

What does the update mean for PR strategies? 

The update emphasises the importance of product PR as part of your approach.

If you’re not already, incorporating product placements, samples and reviews into your always-on activity will be key in gaining visibility in AI search. 

While backlinks were once used to judge website performance, LLMs are prioritising context and relevance of brand mentions, making product PR the perfect tactic. 

To support this, expert commentary remains an important trust signal so balancing this with product PR will be key in building an authentic, trustworthy brand in the eyes of LLMs. 

YouTube are Testing AI Overviews in Search Results

Today, it’s been reported that YouTube are starting to test AI overviews within their search results, in order to highlight what they deem “the most relevant clips” for individual users; based upon their search.

We predict that this will not only impact user interactivity with watching longer videos in full, but may also cause a shift in search behaviour from within the YouTube platform. In theory, by delivering this type of snapshot video content, users will be delivered touristic and local content much quicker, as well as when carrying out product research related searches too.

YouTube will expand the rollout of their AI overviews test based upon the results of a small subset of YouTube Premium members’ feedback, in the US initially.

Google claims “EEAT Can’t be Added to Web Pages”

At a recent conference in New York, Google’s John Mueller claimed that “EEAT is not something SEOs can add to their pages.” He went on to discuss the role of EEAT and stated that adding keywords or metadata to content is, in fact, not what EEAT is about. Mueller emphasised that EEAT is more significant for YMYL (Your Money or Your Life) content, such as finance or healthcare, where accuracy and trustworthiness are paramount.

Non-YMYL content, like blogs, on the other hand, should step away from EEAT signals as they are not a significant factor in Google’s ranking assessments.

So what does this mean for SEO?

Many SEOs base their strategies on hitting as many EEAT signals as possible, but Mueller’s message is that this overlooks the bigger picture. SEOs should instead focus on demonstrating real trust and expertise, not just fabricating it with imaginary industry experts and AI-generated images.

Instead, we should go back to basics: content should be high-quality, informative, and genuinely useful to the user. This is what Google truly values—not web pages that simply tick off a list of points the algorithm supposedly favours. You might think this is basic knowledge, but the recent rise of AI has led many SEOs to abandon strategies that produce content unique, targeted, and personal to the user.

Expert opinion

Cara Dudgeon, Senior Content Writer

"It’s no secret that content is rapidly evolving and that SEO content writers need to keep up with the ever-changing demands of Google. However, one thing that is consistent is that the user is paramount. Sure, EEAT is important in YMYL content, but why stop there? Think about your own purchasing habits. Reviews, expert opinions, and a trustworthy website are all extremely important conversion factors. Users are far more likely to purchase a product or sign up for something that has 1000s of reviews rather than one that doesn’t have any. So, why wouldn’t we add these features to pages? But it’s essential that these features are real and genuine to keep the trust of users that become loyal customers. Fabricating these trust signals with AI can seriously harm your business. Not just in ranking factors, but also in terms of growth and profitability. Negative reviews are just as easy to write as positive ones - make sure your users are leaving reviews for the right reasons."

ChatGPT launches an image generation update

OpenAI’s recent update to ChatGPT, powered by the GPT-4o model, introduces advanced image generation capabilities in the chat interface. This new update allows users to generate AI images in various art styles, including photorealism and stylistic visuals. ChatGPT-4o now understands visuals on a deeper level. It can analyse user-uploaded images, extract content and visual cues, and use the information to create new images. A standout advancement is GPT-4o’s capability to understand highly detailed prompts. While many other AI image generators struggle to handle more than 5 to 8 elements, GPT‑4o can handle up to 10 to 20 different objects.

This opens up countless opportunities for marketers, allowing them to mock up complex visuals such as vibrant product flat-lays with more control than ever before. It also opens up new possibilities for asset creation, especially for those with smaller budgets who are working on evolving visual identities. They can generate, tweak, and refine visuals seamlessly and extremely quickly. The latest model ensures consistency across iterations, which helps build brand recognition and storytelling power.

We’ve seen many users on LinkedIn having fun with this, using the new update to create versions of themselves as dolls. It’ll be interesting to see where this update leads us!

Here is what our senior paid social executive, Emma Sowersby, had to say about this update:

Expert opinion

Emma Sowersby, Senior Paid Social Executive

"This update is transforming ChatGPT from a tool into a creative partner. AI is advancing, and those who don’t embrace the change will be left behind. I believe this update could be significant for smaller businesses, where creative potential is restricted. Moreover, it could serve as an excellent resource for creative ideation, allowing for the creation of mock-ups that can be referenced for design briefs. There’s plenty of opportunity here for marketers to use and learn from. However, on the whole, working with designers to mock up creative assets is preferred. Human designers bring emotion, intuition and strategic thinking that AI simply can’t replicate. AI can mock up images that follow a prompt, but designers have a better understanding of how consumers visually consume images, factoring in flow and storytelling in a way AI never can. Designers not only create beautiful images, but also ones that are purposeful and effective. Another thing to be mindful of is how text may appear on these images (AI-generated images have a history of rendering text oddly), so I would recommend adding any copy afterwards to ensure complete consistency with brand guidelines."

Google Ads Rolls Out a Useful AI Update for Brand Keyword Matching

Keyword Match Types and the logic that goes into matching them to Search Terms has been evolving and changing for years. The saying goes “Exact Match is no longer Exact”, with Google continuously trending toward less precise targeting, instead leaning on machine learning to widen the scope of keyword targeting to include longtail and new search terms that can’t be predicted.

The latest announcement from Google includes 4 key changes that will have PPC advertisers changing the way they work with their brand and the unique search terms in that sphere.

1. Mispelled search terms are becoming easier to manage! If you’ve ever struggled to add negatives of all the potential mispells of a brand term, this AI-driven update means advertisers need only add the core brand term. AI will ensure that mispells are included.
2. Brand Exclusions across ALL keywords and DSA campaigns are rolling out, expanding the Performance Max feature to core campaign types.
3. Additional suggestions for Brand keywords and search terms, which turns a Broad Match keyword into a wider brand “sphere” targeting-signal, allowing for “Brand Generic” keywords to filter through more often.
4. Previously un-trackable “Other” search terms will now start showing, giving advertisers more direct control over targeting and exclusions – this last one allows PPCers to “rein in” the expanded targeting that the expansion caters to.

All of this trends towards more consolidation in Google Ads campaign structure, encouraging less fragmentation between Brand and Non-Brand campaigns, joining them up to act as a “Messy Middle Search Campaign”.

Automatically Created Assets Are Coming To Performance Max

Google has started testing its automatically created assets feature on Performance Max campaigns

Performance Max is continuing to be Google’s favourite test-bed, with Google now testing “automatically created assets”. By using advanced AI generative process, this feature takes advertiser’s already-uploaded text & image assets to create something new.

If you’re familiar with Google’s Responsive Display Ads, or Discovery Ads, you’ll understand how this works: Google will “mix and match” text and visual assets together, in order to create visual ads (similar to Display ads) to increase visibility of your ads across networks outside of Search & Shopping. As usual with Performance Max, advertisers can choose to let “the machine” handle decision making here, although how closely these “auto-assets” will adhere to brand guidelines remains to be seen…

Search Generative Experience

Google’s Search Generative Experience (SGE) is finally here!

Well, kind of. On May 25th 2023 Google announced SGE, which added “AI” results to the top of some searches, which is powered by Google Gemini (formerly Bard).

The initial rollout was to users who had opted in and were signed into their Google accounts.

Google I/O Post-Game Update: The Latest Advancements In AI For Google Ads

This year’s Google I/O showcased how the latest advancements in AI can help “boost creativity”, and generally raise the bar for productivity. Google also reaffirmed their objective in building and deploying AI responsibly (we’ll let you decide what to make of that last bit).

Some new products and variants within the Google AI landscape were promoted, such as the new search generative experience (SGE) that will make search engines more of a conversational piece while connecting users to the information, services and products they desire, via “Bard” – this will expand and enhance new and existing apps such as Google Workplace, Maps, Lens and more .