The question of ROI is increasingly popping up in connection with another popular GenAI product: chat. 

Many news organisations are busy building chat products to see if they can open the door to deeper user engagement. 

For example, The San Francisco Chronicle recently launched a chatbot that provides restaurant recommendations after combing through nearly 1,000 different restaurants in the newspaper’s database. The bot has been built off hundreds of reviews in the Chronicle’s Top Restaurant lists, all written and vetted by its food and wine writers and critics, according to Sarah Feldberg, The Chronicle’s deputy director of product and strategy.

Forbes is testing a chat product called Adelaide (named after the wife of founder B.C. Forbes), where the user can input a query and get a summary response as well as a list of related articles.

Skift, a niche travel and hospitality publishing site, does something similar. Ask Skift was trained on 11 years of content.

Germany’s Ippen Media is building a chatbot that appears on the article page and lets the reader ask further questions about the topics mentioned in the story. 

In India, both Times Internet and The Hindustan Times are also experimenting with chat, with Times Internet clearly building a product that would allow it to generate more revenue.

The BBC is also exploring “the potential for chatbots to provide interactive and tailored learning on BBC Bitesize,” its educational service for children (similar to Khan Academy’s Khanmigo effort, perhaps), an initiative that any publisher with a rich trove of evergreen content could emulate. 

But is it worth the effort? This is the question many publishers are asking. It costs money to run GenAI search queries. How does one monetise chatbots?

One option is to place advertisements inside chatbots to generate revenue. It is an option that both Google and Microsoft are exploring. 

Another option would be to initiate the subscription flow from chat itself. 

How about building a chat product that is so powerful, valuable, and useful that it becomes a premium product in itself worth charging for? It could serve your most valuable niche of customers and provide specialised information that you have already published and which your data already shows users are willing to pay for. 

We already have a news-adjacent use case in the legal services bot that Thomson Reuters has launched. Bloomberg also offers chatbots as a service to its clients. 

Could we draw inspiration from these B2B cases? Could we not create a product that draws on publishers’ archives and expertise to create, and perhaps even help us book, an entire vacation, where we specify our budget and the travelers’ interests?

Or one where you can share a photograph of what is inside your fridge and ask the bot for customised recipes that would feed a family of five, where one diner is a senior citizen and two are teenage boys with dietary restrictions?

Or one that helps a young professional draw up a comprehensive list of everything she needs to do ahead of a tax-filing deadline, based on hundreds of personal finance columns and her own financial goals? 

GenAI makes all of this possible, especially as techniques to minimise hallucinations make rapid strides. Most news sites have mediocre search experiences. This could solve that problem — and add in a layer of personalisation, which was a recurring theme at the INMA Media Subscriptions Summit. 

The other gold mine in chat is data. Users willingly engage with the bot and share information about themselves. It’s often worth spending money to gather data (I’ll be writing more about this soon).

Other industries, such as retail or banking, have already discovered that chat fits nicely with a broader commercial strategy, where chat can serve as a tool for lead generation, payments, and retention. Publishers like Rede Gazeta in Brazil are already using chat to supplement the effectiveness of their customer service offering, allowing humans to focus on tougher, more specialised questions that the bot cannot answer. 

All of this translates into money down the line.

If you’d like to subscribe to my bi-weekly newsletter, INMA members can do so here.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .