What AI can do for publishers: Cutting through the hype

Getting your Trinity Audio player ready...

Don’t get me wrong, I’m no Luddite, but whenever technologies are making the critical shift from abstract concept to practical application they seem to go through an awkward adolescent phase where “what it’s called” and the in-depth details of “how it works” seem to be more talked about than the infinitely more useful understanding of “what it does”.

The same thing happened with big data and is undoubtedly on the horizon with Blockchain. I wasn’t the only one who was bemused by start-up HDAC’s recent advertising campaign, aired during TV coverage of the FIFA World Cup and showing a blockchain-enabled house. My response: who cares HOW the fridge and the TV talk to each other, I’m not even that interested that they CAN talk to one another, what I want to know is WHY will their conversation make my life better?

To continue the teenage analogy, technologies that are going through this phase can often seem hard to understand, overly complicated and potentially unstable… not to mention expensive!

Inevitable though this stage seems to be, I think the confusion that it generates can hold organisations back when it comes to establishing the relevance of AI to their business. Nowhere is this truer than in the publishing industry.

In the case of publishing, far from being a solution looking for a problem, AI really does have the potential to transform an industry that is under enormous pressure to stay relevant with consumers and to deliver results for advertisers and partners. I believe that the hype surrounding AI is preventing publishers from identifying the areas where it can simply and cost-effectively solve the challenges they are facing.

Riding the waves of disruption with AI…

Fundamentally, artificial intelligence is a computer programme that we design to incorporate data feedback, and to change in response to that feedback, operating more effectively to achieve the goals we set it. Machine learning is perhaps a better term as it captures the iterative nature of the programme. It’s that responsive, adaptable nature that equips AI programmes to help publishers navigate the disruptive environment they face.

But all that publishers really need to know is that using this type of programme doesn’t need to be expensive and complicated, and that it can solve the critical issue of unlocking the value of under-used editorial. Here’s how:

Propagation, persistence and personalisation…

Publishers invest heavily in generating the original, insightful editorial content that will appeal to their target audience. The best pieces are relevant, timely and build a relationship with the reader that invites them to come back for more. However, with an insatiable 24/7 media cycle, each new piece of editorial content quickly sinks down the agenda, meaning the investment in creating it becomes a sunk cost too.

But the news cycle is just that – a cycle. Before long a related story or topic will start to trend. This is where content aggregation, driven by artificial intelligence that identifies trending stories and the existing stories that are associated with them, allows that original content to resurface and propagate alongside the newly trending story. This related content provides much-needed context for users who are facing information overload and strengthens the customer-publisher relationship by adding value.

For publishers with multiple media properties this tactic can work across them all. By categorising stories into themes and topics, a great article in one title can cross-pollinate to another title in the portfolio when the story becomes relevant for that audience. This means that content persists in adding value long after its original publication date and is served at the right time and in the right place to maximise user engagement. This is definitely the kind of intelligence – artificial or otherwise – that helps publishers get better return on their investment!

AI also allows publishers to understand their audience at a much deeper level. By analysing the mass of data created when users interact with their sites and cross-referencing with user’s stated preferences when they register, publishers can develop detailed profiles so they can automatically serve content and ads that are highly likely to appeal to users. Not only does this significantly improve site monetisation, but automation is a huge bonus for businesses that are operating with lean staffing levels – it really is a labour-saving way of developing an efficient, high-quality product that is tightly tailored to audience interests. Personalisation is the all-conquering panacea for consumers right now and AI helps publishers respond to this demand to create sticky sites that keep users engaged.

Of course, these days, everyone is a publisher and the beauty of social content is that it can be harnessed by publishers to augment their own bespoke content. Drawing in wider socially generated content keeps sites well-fed and broadens their appeal for users who want to get their fix of social trends, news and opinion from a single source.

So, while AI is unquestionably more than a fad, it is definitely going through a phase. Once we look beyond the hype I believe that publishers can turn adolescent angst into teenage kicks and start to benefit from the useful and accessible ways machine learning can help them overcome the challenges of disruption.

by Ashwin Saddul, Managing director at Better than Paper.

—

Re-published by kind permission of FIPP, the network for global media