United Robots, a Swedish company, creates thousands of pieces of automated content every day. It’s one of a new breed of businesses helping publishers and marketers to automate their content creation.
We spoke to Sören Karlsson, United Robots’ CEO, about the power and potential of content automation, for one of a new series of podcasts.
What is content automation?
Content automation is creating publishable and readable text from large datasets.
We’re using a technology called natural language generation (NLG). It’s a fairly old technique, but it has become really useful in the last five to ten years because of the big data sets that are everywhere and because the cost and power of computing has changed.
Can you give us an example?
We’re working with almost all local Swedish newspaper companies. Local online news companies have been in retreat and can’t afford to cover local events anymore, but they can automate it and keep covering those events in a much more cost-effective way.
It’s straightforward to automate sports texts. The data is really structured and easy to analyse. There is a lot of local football played at all kinds of levels in Sweden. All the results are reported into a database and our algorithms immediately analyse them.
We are looking for journalistic angles in the data. Has something unusual happened? Was there an especially big win? Does this result mean one of the teams is leading the league?
Then we produce the text which is published immediately on the right website. In less than a second.
What volume of content is being automated?
We produce thousands of texts every day. One of our customers is called MittMedia, which is the largest local news publisher in Sweden. They publish hundreds of sports texts every day. We also automate stories about real estate sales. We write over 100 such stories a day for them. So that’s the kind of volume we’re talking about.
What are the results – more traffic, people staying longer on the page or conversions?
All of that. It’s really good if you want to convert free riders into paying users.
One thing we now see a lot of is personalisation. If you have readers who are logged in, then you know who they are, where they live, what they’re interested in and then you can pinpoint them with exactly right local content.
What can content marketers learn from what publishers are doing through content automation?
Sit down and look at what kind of data you have. Can this data tell stories that are valuable for you as a company or for your customers or users?
If the answer is yes, then automation is an option to examine.
You need a lot of data to get the return on investment on this technique. You need to be able to produce a lot of text and you need to be able to distribute it in a meaningful way.
How can brands maintain quality when they’re automating content?
You can’t just leave it to the machine.
The best results happen when man and machine work together. You need to have your people involved in producing these texts so you can tweak the robot, so it feels like you’re writing these texts.
What does the future look like for content automation?
Ask yourself some simple questions: Will data sets continue to grow? Will data centres continue to get faster and cheaper? Will the demand for personalized information grow? I would say yes.
So you have a bright future for even more content automation. I think we have just seen the beginning.Contact us