How Sage, A Global Media Publisher, Uses AI

“I love conversations like these because they make AI real,” said Jasper VP of Marketing Meghan Keaney Anderson. “We talk about AI a lot in terms of concepts —the big promise of it — but it only really matters when you get into the brass tacks of how this is working in your day to day.”

The conversation Anderson is referring to is a live panel discussion — “How Large Media Companies & Publishers use AI” — between herself and Shellie Johnson, director of global marketing at Sage, a global academic publisher of books, journals, and library materials.

Johnson is responsible for managing the martech stack and operations for Sage’s global marketing team with multiple divisions. She helped coordinate AI’s adoption, specifically Jasper, across Sage’s marketing org as her team ran campaigns with emails, digital ads, webinars, and conferences.

In the discussion, the leader walked Anderson and the viewers through Jasper’s brass tacks, day to day use at a publishing company. The conversation is especially significant considering the unique risks and opportunities with AI adoption at an international publisher of educational materials — transparency and trustworthiness being among the biggest.

Johnson broke down the why, the how, and the use cases behind Sage’s adoption of Jasper by touching on Sage’s multi-disciplinary generative AI committee and the role it played in offering guidance and support. Johnson also walked listeners through Sage’s Jasper pilot program and how team members who were initially fearful of AI making them obsolete changed their opinions once they actually used the tool.

There’s even more great insight that Johnson delivered on Sage’s team-wide experience using AI for marketing. Hear what she had to say by listening to the full 32-minute conversation or reading through the transcript we provided below!

{{rule}}

*This transcript has been edited for length and clarity

Anderson: What were some of the problems that you were trying to address when you decided to explore AI?

Johnson: We're always looking at how we can become more efficient. When ChatGPT became mainstream and everyone was excited about it, our company took the initiative for us to explore. We were trying to see how it would fit within marketing. The exciting thing about ChatGPT is that you can do a lot of experimentation at an individual level, but what we really needed was something more secure to protect our proprietary information.

We also needed something a lot more structured. We needed an interface that could be incredibly organized where teams could work together and share.

Anderson: You have a generative AI committee inside Sage. Can you tell me a bit more about the committee?

Johnson: We were encouraged to explore [AI] freely on our own, but we recognized that, especially for protecting our proprietary information, we did need some guardrails. So it made sense to form a company-wide generative AI committee.

It’s meant to be more of an advisory board. We provide policies and guidelines, focused around being ethical, legal, and secure using generative AI. We also recommend AI tools, vendors, and partners. We make sure that we understand the tech infrastructure and capabilities. We share our learning and develop best practices. We also support training and we make sure we communicate all that across the business.

Anderson: One of the decisions that you have to make in this process is: Are you going to build or versus buy? Are you going to use an underlying foundation model and build it into your systems internally? Or are you going to select a co-pilot that is more end-to-end and purpose-built for a specific use case, like Jasper. You obviously went with the latter. Can you share a little bit about that decision process?

Johnson: Corwin, our K-12 subsidiary, actually got in touch with Jasper first. We decided to broaden it across multiple marketing divisions in a pilot with Jasper. for over a month. We had key people from the various marketing divisions participate — mainly our marketing communications departments because that's where we thought we would see a lot of strong use cases.

They were actually comparing what they would get out of Jasper against ChatGPT. They were really impressed with Jasper. They liked the templates and that [the platform] can be organized in a way that the team could use it interchangeably. So it made it pretty easy to put our proposal forward to go with Jasper across all of our marketing divisions.

{{blog-quote}}

Anderson: Well, that's really nice to hear! What were some of the early use cases that started to gain traction?

Johnson: I have a pretty heavy use case that Jasper worked so closely with us on, and this is just phase one of it. We have a marketing communications team in our U.S. college division, which focuses a lot on textbooks.

They do copy development for over 100 new textbooks every year. They have been using a simple word document, what they call a Product Information Page (PIP). That's where they include different product descriptions like unique value proposition, key features, and more. These descriptions are then fed into our website and multiple campaign channels, including email, social media, and paid ads. Typically they start by reviewing what we call the Author Prepared Preface for the text and they create the necessary options based on that information. Distilling that takes a lot of time. And having over 100 titles a year obviously compounds that.

So they worked with the Jasper team to develop a custom workflow for this creation. Step one: they input the title, author, and preface into Jasper. Within seconds, it outputs drafts of the basic description that they need. Then they can easily edit and save them over time. Step two: get Jasper to generate effective keywords to boost their SEO based on the product descriptions in step one.

That's what they're doing right now and already seen it's saving them an enormous amount of time. They're working with the Jasper team to fine tune this even more to help them develop copy for their Amazon Onyx descriptions, social media, and the author marketing tools in the same process. They foresee that this will reduce the time spent on PIPs, it'll boost SEO, and it'll just create more engaging content. We can see how this methodology could work in other scenarios as well.

Anderson: I love that level of detail, especially with this group of people because everybody on this call knows that pain. What are the ways you are using Jasper to optimize your content for different formats in addition to just writing things faster?

Johnson: SEO is a great example. We have a digital ads program. Some of our divisions have already moved over to a new web CMS tool. So SEO is on the forefront [of that work] and we’re definitely using Jasper for that in particular.

I also keep hearing great examples from webinars and that it's not about the ideation of a webinar or the marketing content to promote it. We're looking at use cases after the webinar, where you get all the audience questions and typically marketing isn't the subject matter expert. So one use case we want to implement is putting enough information in the Knowledge Base in Jasper to have a quicker turnaround to answer these questions because right now we're relying heavily on editorial and that could take up to a week or two…it should uplift our content and prove more engagement with our audience.

Anderson: I really love that. Let's get meta with it for this webinar.

*Here, Anderson suggests taking the transcript of the webinar, including all questions from the chat, to create a blog post that addresses more insights and learning on the intersection of AI with the media and publishing industries. So here we are, getting meta!

Anderson: Let's talk about adoption…Walk us through the process of beginning to adopt Jasper or any AI into your team. What were some of the challenges that you had to overcome? And how did you think about rolling this out?

Johnson: With the pilot, it was somewhat easy because we made sure that we had individuals participating who would use it heavily in their day to day. Once we signed on with Jasper and rolled it out to the larger group, that's where we were hitting some roadblocks.

I think a lot of it had to do with fear. People are fearful that this could actually take over their job. If you don't go into the tool, I can understand why you're fearful because as a marketer, you're used to writing your own copy. And if you're not writing your own copy, what would that mean for your job?

So we wanted to make sure that we could highlight use cases where it wasn't about an AI tool taking over for you — it’s an AI tool that is a copilot helping you with the ideation. It’s fine-tuning copy that you created in multiple ways so you could actually focus on more of the strategy or other things that you never have time to do.

What we really want to test now is [whether] a marketing manager sees value in it and in their day to day. I'm a little removed from it. I don't write copy all day. So we assigned someone in our customer strategy group — our comms manager. He's now the lead for Jasper and collecting use cases. He's going to promote these use cases heavily and widely across our marketing divisions. He's also putting together metrics to show how the tool is improving our content, improving efficiencies, and where there could be cost savings.

So we want to move it from individuals or small teams to wider use and how they can effectively use it in their day to day.

Anderson: That’s so interesting. I think one of the best ways to attack that inherent fear is to find the thing that would be of value to the end user; help them get busy work out of the way and enable them to invest their time in the things that they care about. And so much of it is about getting down to that use case. Once that use case is found, then the fear melts away because you start to see your place in it. You start to see, “I can use it for this [small] thing over here, but not this grander thing that I'm very necessary for.”

That is not unique to publishing or media. We've seen that across all of our industries. Like anything that is new and scary, the best way to get over that fear is to understand it better. Understand your place in it and the way that you can wield it to make your life or your outcomes better. So that's a great example.

Another thing that you have to do: teach everybody as you’re onboarding, and establish your company standards for responsible use. Especially in publishing, many of you are dealing with educational content. It’s high stakes material.

So how do you think about setting ethical standards and rolling out practices of responsible use of AI in the business?

Johnson: Company wide, not just how we're using Jasper, we look at it as “You have to have a human in the loop.” So since quality and accuracy of our content is essential, we're primarily using AI to help us draft content that undergoes rigorous human review. This is outside of marketing too because we stand by the quality of our content.

So it's because of this non-trivial task that best practices should include transparency if we're publishing content fully generated by AI. Whether it's assisted by AI or generated by AI on the fly, readers and users know the risk of inaccuracies and bias. So it's about being transparent and making sure that there are some guidelines that we're aware of to make sure that we retain integrity as a publisher.

Anderson: Are there types of content that you don't use AI for intentionally? Or is the approach more about transparency?

Johnson: We're still evolving that right now. That’s why the [AI] committee makes perfect sense so we can have these use cases run by the committee. And obviously each case is different and unique. And we want to make sure that, going back to the transparency of our researchers and our authors, we’re consistent in that message.

Anderson: I love that. Because so much is changing and evolving, having that central committee to be the clearinghouse of absorbing those questions and making informed decisions about them for the company is great.

What's the composition of that committee? What parts of the company are on it? People may be listening trying to figure out if they should spin up their own and that would be really constructive.

Johnson: A lot of it is tech heavy. We have a lot of great representation from our technology side of the business. We have product innovation. We have a data scientist on board. And legal.

Anderson: Great. You said that throughout the pilot, and beyond, you've been tracking results to understand what's working and what's not with your use cases. Can you share some of what you've seen, in terms of impact?

Johnson: A lot of it is around time savings. There are a couple of use cases where we see potential cost savings as well; one is around translations.

We have one group that, years ago, outsourced translations and it proved to be cost prohibitive. So the only time that they would do translations is if they had internal resources within marketing that could do those translations themselves. That was very manual and would take a lot of time.

Now that we have Jasper, we're doing translations again [specifically emails to APAC countries] and it takes 15 minutes.

Learn how you can take advantage of this technology with A Marketing Team's Guide to Generative AI‍

Anderson: I'm nodding energetically because in a prior role at a global company, we had an educational arm. We had a localization process into multiple languages and that process was a month long. So for any educational content, you would have to plan in a month of translation before you could actually hit students or hit the market. It was, to your point, eye-wateringly expensive. So I definitely feel that pain and love that use case.

Johnson: In another one, Corwin, our K-12 market, would outsource copywriters to write their back [book] covers. And it would cost around $30,000 dollars a year. And now they've been able to bring it back in house. Jasper can do it for them. So that's exciting.

Anderson: Do you still have people inside doing the quality control?

Johnson: Yes.

Anderson: Great. That's awesome.

*There’s a little admin discussion about addressing questions from viewers before Anderson jumps into a future-forward question.

Anderson: Where do you think AI in the publishing industry is going to head?

Johnson: I cheated a little and I actually asked our VP of Books and Social Science Innovation to comment on this. She's the co-chair of our generative AI committee.

She said the way she'd like to see it evolve is, “To become a powerful tool to augment human creativity to allow us to work on more meaningful tasks and to help us to think better. There seem to be many obvious opportunities for improving processes in higher education publishing using generative AI: copyediting; proofing; creating what we call Mark Records; alt text and long descriptions for accessibility; and generating ancillary materials like revision questions. That will mean our authors can focus their creativity, talent, and energy on the core ideas and content of their work.”

Not my words, but really good!

Anderson: I love that. I'll ask you a more practical question: What would you like Jasper to do for you? If there's a feature that Jasper doesn't have that you would love to see us develop or a use case what are your asks for us?

Johnson: I oversee our martech stack right now. We have gaps in our technology around managing content and digital assets. And it would be amazing to be in one tool. I see the potential for Jasper to fulfill our content management needs using Campaigns, especially now that you have Spaces, so each of the teams can create their own libraries using that functionality.

Anderson: Perfect. We'll take that back to the product team. They're always looking for new ideas.

There were a couple questions that came in over the last few minutes. What was the biggest learning curve for your team as you started to adopt [Jasper]?

Johnson: Again, I would say people being fearful of using it. We capitalized off of those that jumped right in and saw the benefit of it. They would highlight what they were doing and that helped to highlight those specific examples.

Anderson: Yeah that makes a ton of sense.

I think what I admire about Sage and what you all are doing is you're finding a way to balance being an early adopter, being pioneering in the space, and not sleeping on innovation in, but doing so in a way that is thoughtful, responsible, and really puts the quality of the content at the at the forefront.

And so often we feel like those are at odds. You either move fast and break things or you’re conscientious and you move slowly. It seems like Sage has found a way to move pretty quickly to heavy adoption. So kudos for you and your team for diving in and doing so in a way that is thoughtful and scalable for the long term and that puts quality content at the heart.

*Here, the webinar wraps up! Thanks a lot for making it this far. Again, if you’re interested in seeing the full webinar, follow the link here. Happy creating!

‍