This week, we held our third ‘AI Special — Q&A’ webinar, in which the twelve contributors to the AI Special we ran in the September / October issue of InPublishing magazine answered questions.
There were lots of useful insights — I have picked one from each of the panelists below:
- Site AI: Consumer expectations are changing; when somebody comes to a website, they are looking to consume content the way they want. They might want to read, to interrogate, to listen or to watch. Users will want to select the mode that they consume content in. (H/t Markus Karlsson @ Affino)
- Upskilling: General business is an area of massive growth potential when it comes to using AI. The task is to upskill everybody in terms of their day-to-day work and what AI can do for them. We need to educate everyone, to get them over the barriers of ‘where do I start?’ or ‘what do I do?’. (H/t Terry Hornsby @ Reach)
- Unlocking your archives: A publisher’s extensive archives give them the golden opportunity to reclaim some of the authority that magazines have traditionally had by applying LLM interfaces to their own content. (H/t Hutch Hicken @ BlueToad)
- Content discoverability: AI can greatly improve discoverability by assisting at the content production stage: categorising content, tagging it in the right way, analysing data, and using those learnings to drive discoverability. (H/t Beth Ashton @ Bright Sites)
- Hack days: The start point for running a successful hack day is to begin with a long list of problems that you think need to be solved and then whittle it down well ahead of the day so that you’re going in with a good mix of tangible customer facing and business facing problems. (H/t Jenni Allen @ Which?)
- Clean data: To get the most out of AI, you need to make sure your data is up to scratch: clean, dupe free, up-to-date and not spread across multiple systems. A single source of truth for the agent or LLM to read will deliver the best results. (H/t Rob Green @ ESco)
- Content strategy: Intelligent tagging of content allows you to properly analyse what you currently publish and what you might want to publish in future. But… journalists aren’t always the best taggers! Properly instructed, AI does it better! (H/t Adriana Whiteley @ FT Strategies)
- Spotting patterns in the audience journey: People who end up either subscribing or taking up an advertiser’s products or services, don't behave like butterflies. They circle around a problem, they come back a few times, they read something practical, maybe watch a video, maybe do a webinar. AI is particularly good at spotting that repeated sort of dance. (H/t Damian McAlonan @ Mark Allen Group)
- Adapting content: Where AI is really starting to prove transformative is in helping publishers adapt content across different formats and different platforms. So, as an example, a single article can now be repurposed into an audio bulletin, a social media post, a podcast script, or even a video summary. AI is doing the heavy lifting, but the human remains in control. (H/t Martin Ashplant @ PA Media)
- Efficiency savings in image editing: By adopting batch automation for mass image editing and correction, thousands of images can be processed on a daily or weekly basis. It’s very quick and very accurate. (H/t Derek Milne @ Pixometry)
- Training an AI model: People often say if you're using AI at work, treat it like an intern that you're going to give a task to – you're going to check it carefully when it comes back before you use it. When it comes to training a model, it's less like an intern and more like a naughty puppy. Just because you tell it to do something doesn't mean it's actually going to do it! (H/t Carl Myers @ Faversham House)
- Rethinking workflows: Use the momentum which exists around AI to rethink the way you work. While you’re exploring the potential of AI, you can quickly uncover various inefficiencies in your workflows. It's important to keep an open mind and actually choose the right fix for the problem. Not every solution or issue requires an AI hammer, for example. Sometimes the best solution to a problem might be something else in your toolbox altogether, like, for example, a process change. (H/t Tom Pijsel @ WoodWing)
In ninety minutes of in-depth Q&A, there was loads more advice and insight, which you can catch by watching the recording here. (You will need to fill out the registration form and then you’ll be directed to the recording.)
