Ask my site: Boosting media UX with generative AI search.

Dear fellow media pros, don't just see Google's generative AI search as a threat to your audience development. The promise of a conversational search for users can radically change their on-site experience: an opportunity not to be missed.

When Google announced the beta version of its?generative search engine?in May, it was met with considerable concern in the media.?For years, Google's practice of aggregating information and presenting it in search results without requiring users to visit the original sources has led to a loss of traffic and advertising revenue. Now, with Google's AI Integrated Search, this issue appears to have moved to the next level.

The search engine allows users to engage in natural language interactions,?providing well-structured, conversational answers with verified sources. This level of detail and quality in search results could potentially satisfy users to the extent that they no longer feel the need to click on the links presented in the top right-hand corner of the demo.

From a user's perspective, however, this could be the most significant innovation in the search engine space.?It represents a radically new approach, introducing a dialogic search experience in which the engine itself suggests step-by-step queries, leading the visitor to a comprehensive and personalised exploration of topics. This discovery not only includes established media sources but also highlights alternative perspectives from niche websites, individual content creators, and Reddit groups.

Three reasons to look at generative AI search differently.

First: AI generative search can strengthen your site as a destination.

Today more than ever before, the objective of a media outlet is to break free from the forms of traffic driven by others?- social, above all - and become a destination point that gives its visitors as many reasons as possible to return and consume content.

As the NYT Ceo?Meredith Kopit Levien recently recalled:

“Our job is, number one, to obsessively focus on getting people to come to our destination and build a direct relationship with us, to register with us, to give us their email address, and to let us show up in their inbox”.?This is the ultimate tool for mitigating the risk of disintermediation resulting from new interfaces.

Second: onsite search has been neglected.

Of the many features and offerings media companies have developed in recent years, also using AI - recommendation engines, curated lists, newsletters, alerts - internal search has remained stuck in the 1.0 version. This has been a priority more for media specialising in financial or academic content, or big media aggregators (mostly for brand TV program searches).

However, active search is a primary behaviour of our digital activity,?even making a comeback on platforms where you would not expect it (NYT:?“For Gen Z, TikTok Is the New Search Engine”)

Also, consumers often complain about the "paradox of choice" or their inability to find something interesting to watch on streaming platforms that have incredibly large libraries of on-demand content.

These same users have been self-educating themselves over the past six months all over the world to write prompts for ChatGPT, for work, or just out of curiosity.?OpenAI's bot has created an expanded user base ready to interrogate any system from which they can extract useful and desired information.

Third: it is technology you can use, too.

The barriers to developing advanced generative search are now at their lowest.?Google, while focusing on its search engine, also offers out-of-the-box enterprise search based on its?AI App Builder: Organisations can create custom chatbots and semantic search applications in minutes, with minimal coding required to get started, the ability to handle multimodal data such as images and control over how to answer summaries are generated.

For those who are not interested in going to Big Tech, open-source Gen AI is a reliable and growing alternative.?PerplexityAI?offers an 'interactive AI search companion' called Copilot, which provides detailed filters before searching the web and presenting results in a ChatGPT-like experience. Its interface includes real-time citations to online sources and follow-up questions.?Kagi?offers many features including AI that can summarise the entire results list, summarise individual pages in the results list and answer follow-up questions on a returned result. More to come.

Which means:

Time to give an AI search power boost to your site.

Infinite use cases.?Imagine giving your visitors and subscribers the ability to request a summary of recommendations for places to visit this weekend, depending on distance and weather (from your Weekenders section), or a comprehensive explanation of the roots of the Russian war against Ukraine (from your Past Archive, your Op-Eds, your articles from the last 30 years). Or a list of companies to consider for bond investments based on analyst recommendations over the last six months. Or readers asking for a health programme for the over 50s (based on your health and fitness archive).

The opportunity is to turn your past content into utility: one of the biggest levers for retention?(again, NYT or Die Zeit in Germany are good examples. Potentially, data enthusiasts and business users could also ask for on-the-fly charts summarising economic, financial, or demographic data.?The above examples are just the tip of the iceberg and are already achievable?with the available language models and interfaces.

Ask Skift.

Skift is a media company and platform that focuses on providing news, insights, and research about the travel and tourism industry.

In May 2023, they launched?Ask Skift, their AI chatbot search engine dedicated to the travel industry. They introduced it with an exemplary list of questions it answers: "How does Airbnb plan to use AI?" Or "Who is the new CEO of IHG?" Or "Who owns Ace Hotel?" Or "Write me a short essay on the state of tourism post-Covid."

Ask Skift, which is free for all visitors with up to three questions per month and unlimited for paying media subscribers, is built on top of OpenAI's GPT-3.5. The next step will be to learn from user queries what features to add, and then decide whether to move to OpenAI GPT-4 or an alternative LLM.

The big lesson from this launch: don't wait, experiment together with your audience.?And be patient: the market for off-the-shelf or bespoke solutions will mushroom, so a limited investment is fine.

Conclusion.

The adoption of generative AI search engines should not be seen solely as a Big Tech threat to your traffic but as an opportunity to strengthen the relationship with your audience.

The possibilities to leverage generative AI search are endless.

By enabling visitors to request personalised recommendations, explore in-depth topics, access historical archives, or get real-time data summaries, media sites can turn their content into a utility and improve user retention or offer new reasons to convert to paying subscribers.

A key takeaway from the Ask Sift example is that, as the market for off-the-shelf or customised solutions continues to grow, a limited investment now to explore these technologies can yield significant benefits.

Republished from my blog: thegoodattention.substack.com. To get the next posts in your inbox, you can subscribe here.