The path of progress

The acronym “GPT” in ChatGPT stands for Generative Pretrained Transformer. A Transformer is a type of neural network originally developed by the Google Brain team in 2017. It substantially advanced generative Al models and became the go-to solution for the next few years.

In 2021, Google released research on LaMDA, an advanced transformer-based architecture that was specifically trained on human dialogue or conversation.

LaMDA conversed with users so convincingly that a Google Al ethicist was persuaded the model possessed sentience or conscious feeling.

18 months later OpenAl unveiled ChatGPT, also built on the Transformer architecture and fine-tuned for human conversation. Its seamless user experience resulted in quick adoption, reaching 100 million users in January 2023.

Since then, Al in search has taken center stage. After Microsoft incorporated ChatGPT into Bing, Google promptly launched Bard in February 2023, integrating its internal LaMDA AI model with its search functionalities.

One question naturally arises – how will these enhancements in Generative Al impact paid and organic search differently?

How do recent Al advances affect paid search?

The use of Al in paid search advertising is not new. Google has long since incorporated its Al capabilities into its Google Marketing Platform stack, to help with producing or refining ad copy.

Take Broad Match for example – a keyword match type that allows Google to automatically expand the breadth of search terms where advertisers appear.

In 2021, Google enhanced Broad Match using the large language model BERT, the predecessor to neural network architectures such as GPT and LaMDA.

BERT can extract intent from language and can attribute semantic significance to keywords. Following the integration of BERT with its search engine in 2020, it was a logical next step for Google to make its capabilities available to advertisers.

Google also advocated for the adoption of Responsive Search Ads (RSA). Given a collection of search headlines and bodies, RSA automatically identifies the best combination and creates the most relevant ad to match a customer’s search terms.

Dynamic Search Ads (DSA), another Google feature that has been available for many years, now largely depends on Al models. DSA uses large language models to comprehend the content of landing pages and ad copy.

This enables it to create search ad headlines for advertisers with well-developed sites or a large inventory. These can augment keyword-based search campaigns by filling in coverage gaps.

The integration of Al into the current paid search landscape is broader still. Performance Max campaign functionality lets advertisers combine the above with Al-driven attribution, optimization, audience modeling, and more.

How do recent Al advances affect Organic Search?

Google’s position on SEO is less clear than its commitment to Al across Google Ads. Google has provided guidance around Al-generated content, summarized by this key phrase:

“Our focus on the quality of content, rather than how content is produced, is a useful guide that has helped us deliver reliable, high-quality results to users for years.”

Google recognizes there is a clear playing field of fair use of Al-generated content. However, Google also outlines the risk of using AI-generated content for the sole purpose of manipulating search ranking results. In that context, it points to “SpamBrain” which is Google’s Al-based spam-prevention system.

Using the technology as an advanced language assistant (e.g., to rephrase your self-written content) will not be penalized by Google, but there are use cases that will be.

Technology that detects whether a piece of text was authored by a human or an algorithm is already available.

Additionally, OpenAl is studying how cryptographic watermarks can be embedded in generated text, enabling other services to differentiate it from copy written by humans.

Given the advanced Al capabilities of Google’s SpamBrain, it’s likely that detection strategies will be integrated as a signal to detect spam or malicious content.

How are advertisers impacted by a chat-based search paradigm shift?

What if conversational Al becomes the primary search method? This is an important question to ponder.

How would this shift impact publishers and advertisers? Bing displays sources within the generated answer, yet it’s unclear whether users click on them as often as a typical search interface. This could have a notable effect on the amount of traffic for publishers and the reach of advertisers.

Microsoft CEO, Satya Nadella, attempted to reassure advertisers by asserting that the success of the Chat-Search integration hinges on its ability to generate traffic to the publisher’s website.

He acknowledged that this outcome is crucial because otherwise, publishers may begin blocking Bing’s web crawlers from accessing their content.

Uncertainties remain, however. Could advertisers purchase outbound linked source references in the answers to specific questions? Will brands be willing to pay for their names to be included in responses?

While these questions remain unanswered, the new search paradigm also presents opportunities.

Personalization is increasingly important, and the chat interface offers more personalized user engagement. This could allow Microsoft and Google to collect more accurate data, resulting in more precise and relevant ad experiences – not insignificant given the looming retirement of third-party cookies!

We may also see a more competitive search landscape, resulting in lower costs and superior products.

Key Takeaways

  • Al has existed for years – established players and new entrants are making new leaps in its capabilities.
  • A major shift in consumer search behavior could be near after ChatGPT hit 100 million users – one of the fastest-growing interactive applications in history.
  • Google and Microsoft have built on similar Al functionality to innovate platform capabilities, features, and products.
  • If you plan to use generative Al for content creation to drive visibility in Organic Search, be sure to use it appropriately, understand limitations, and do not share any confidential information in an open-source platform like ChatGPT.
  • Embrace change, experiment, test, and learn to take advantage of everything Al can do.
  • Al cannot do it alone. Al-empowered marketers are an inimitable force.

From strategy and planning to activation and optimization, our teams have been trained on and empowered by the latest advances in Al to drive value for our clients. At KINESSO, we are incredibly proud to be at the forefront of the Al revolution and are set to keep leading the way with our clients.