Partner content

This feature has been written in partnership with Bridged Media. It can be challenging to visualise "Operationalising AI" and what that means for your organisation in terms of people, process and tech. Bridged Media is currently running a survey to help establish an industry benchmark on how to future proof your business using AI enabled tech.

Take the survey to help find out if your newsroom is AI ready. Or go to bridged.media to find out more about how you can optimise content, boost audience engagement, and maximise revenue with AI.

Bridged Media’s Maanas Mediratta on why ROI is increasingly important in AI

For publishers, the conversation around AI has moved on from how to incorporate it into the business. Now, ‘fear of missing out’ has been replaced by ‘fear of not being able to measure’.

This is the latest in our Media Briefs series of short, sharp sponsored episodes with a senior executive from a vendor working with publishers to make their businesses better.

In this episode, we hear again from Bridged Media CEO Maanas Mediratta. Last time we spoke about how Bridged Media was making AI tools more accessible to a broader range of publishers. This time, our conversation was focused on how publisher conversations around AI have moved on, and why measuring ROI is becoming a key part of the discourse.

PH: When we spoke before it was about making AI accessible to more publishers. Are we at a different stage in the industry now, where people are asking how they can measure the impact of AI on their business?

MM: I think 2023 and 2024 was more ‘Hey, we need to incorporate AI.’ Now that FOMO has gone away, or at least has moved to a different fear, which is the fear of not being able to measure.

Post two years of the hype cycle in this innovation, Gen AI, it was natural that more and more people question the tangible value. We’ve seen that consistently across everyone that we interact with, whether we work with them or not. They’re leading with what is the tangible value, rather than what does this product do?

PH: So, how much money is this going to make or how much money is this going to save, or even how much money is this going to cost?

MM: Yeah, and I think this is something that you would witness with any sort of innovation. Slowly, steadily, the ROI comes in.

At the end of the day, no business would want to keep investing until it actually creates something tangible for them. While the wow factor is there with Gen AI, the conversations are now, what is it exactly in the business that’s going to get boosted? Or what can I save, things that directly impact the business?

PH: Is it difficult to actually put a value on AI investment?

MM: It is difficult because, usually, Gen AI touches multiple functions.

A product person is supposed to control how the product is being ingested with this new gen AI tool, whichever tool you might be using. Then the editors, people who are the gatekeepers of content, want some say. Then, of course, it needs to touch the commercial part.

Being able to measure ROI is difficult; being able to measure it in a way that decides whether you want to keep investing or not, is even more difficult, because then there are more stakeholders.

Then it becomes about strategic decisions, what would be the future in terms of the money it can generate or save?

PH: Is there an easy way to put the pieces in place to help with ROI measurement?

MM: Whenever I talk to someone in the media landscape, I always tell them that the first angle, or the first view that you should have, is how can AI enhance my current value proposition?

Different media companies have very different value propositions from being the fastest to report something to having a very niche opinion or view on certain subjects.

Being able to map out the enhancements you can make to your current value proposition is step one. Step two is being able to break it down into the key goal that it is accomplishing. If it is saving time, what is the exact KPI? Is it the number of content pieces per person? Is it the additional impressions that are being created with AI enhanced content? Creating additional engagement, additional loyalty.

Really, being able to come up with a North Star methodology is a key element. That North Star should be hand in hand with your current value proposition, your current value to the market. If you’re able to honestly answer both questions and come down to that North Star, it will definitely be the first big step to making your AI investment tangible and ROI friendly.

PH: Do the people you are working with have that North Star?

MM: I’ve seen AI maturity at different levels for different media companies. There are companies that are still at the stage of, ‘Hey, all our competition is doing something. We have to do something. Let’s just invest and then find tangibility or ROI later’. This group is becoming way smaller than in 2023 and 2024.

There is another set of people, early in their experimentation, where they have found certain pockets of value. These pockets of value, whatever the use case, fall under two umbrellas – enhancing productivity or enhancing customer experience, depending on their internal priorities.

A third, where they have some data, are able to follow some sort of North Star KPIs and are mature in terms of their AI adoption.

So, no real trend, but three kinds of publishers with different maturity levels in measuring ROI.

PH: Is it about mapping traditional KPIs across to AI implementation?

MM: I think at the end of the day, these North Star KPIs have to reflect what the business wants to achieve. They might still be traditional, but that doesn’t mean that it is a roadblock for their AI adoption.

A lot of publishers are still stuck in that ‘impressions are my key North Star’ KPI. That is changing, especially when they’re seeing search volume is not predictable and programmatic revenue is declining. They are really looking to create better audience focused KPIs.

What I always tell them is that there is no right or wrong KPI, as long as you’re able to set up these processes, set up these systems, where you can measure in an easy way what’s actually enhancing your value proposition.

PH: What would you recommend to publishers looking to put measures in place?

MM: I think one of the key things that publishers have to do is be as agile as this ecosystem is.

Unless they are confident of a particular KPI and they are able to measure it properly, they should go with what the industry would call a lean startup cycle. Think about creating a KPI, think about a prototype, whether it’s a solution to enhance your content generation capability, engagement capability, revenue generation capability. Those are the three key buckets that I see.

Was I able to measure the KPI? What was the struggle? Was it an internal alignment struggle? Was it a system struggle? Was it data that we are not able to measure?

Then go in small sprints to refine the model. Whatever implementation you do, you have to be able to measure it. If not, you’re in that quick sand. And believe it or not, time is ticking.

PH: Are certain AI implementations seeing better ROI?

MM: Yeah, the two categories that I mentioned: productivity and customer experience. Then, come back to your core value proposition to the market.

I would really focus on productivity and then go deeper. Productivity has an overall content cycle – discovery, generation, drafting, processing to make it visible, and then analysing what is doing well and what is not doing well. If you look at the overall content cycle, publishers who have that focus on productivity, they need to double click on which part of the cycle they are focusing on to increase productivity.

Connecting the dots to what I said earlier, going with that lean startup methodology, you try something, try to measure it, and then try something else. If you’re not able to measure, then improve on your measurement technique. And if you’ve been able to measure then go to the next iteration, whether you’re building it internally or you’re building with a vendor.

Then on the other side, customer experience, that can really boost a lot of different things once the audience lands. There are different objectives based on the business. Donation based businesses, for example, should focus on engagement because that’s where AI will give you the highest ROI. Affiliate based businesses should be focusing on conversion. How many eyeballs can I create a click with?

Isolate the different processes, whether focusing on productivity or enhancing customer experience, select the category and go deeper – what is a real task that I would like the AI to automate or augment or support. Whichever task you select, think about the KPI, think about your North Star, and then adopt the lean startup cycle. 

Of course, it’s easier said than done.

PH: If you think you’re just going to plug in this AI and everything is wonderful, that’s not the way the world works, right?

MM: Yeah, and I think that’s been one of the biggest reasons for failure that we’ve seen, where people just expect, ‘OK, this is going to just solve my problem’. It’s not. It’s going to take many, many cycles, many iterations. Anything good comes with hard work.

Those failures that you will have while trying to do this AI transformation is the fee that you’re paying towards success. Each failure is making you better. Each failure is making you one step ahead of your competition.

If you’re thinking with one iteration my problems will be solved, I’m sorry, please come back to Earth. That’s not going to happen.

PH: As a vendor supplying AI solutions, do you put an AI ROI framework in place for the publishers that you’re working with?

MM: I think one of our competitive advantages is we’ve done 1,500 AI experiments with publishers across the globe, from publishers that have an audience of more than 500 million a month to niche publishers that are serving an audience of 50,000 or even lower.

We have developed multiple frameworks that can help them understand where they are with their AI maturity and prioritise based on what might move the needle for their business.

A lot of it is done through automated intake forms that we have. We won’t be able to come up with a strategic outlook for your business, but we do provide frameworks that can give hints, especially for mid sized businesses, to know what is the next best action.

PH: Looking long term, do you think the ROI supports ongoing AI investment?

MM: Just sharing results from our business, we have seen within three to six months at least a 300% ROI coming in for the AI agent that we deliver, because it has to.

AI is disrupting a lot of different functions and our principle is that it should not replace jobs. It should augment what people are doing. That’s how we are building agents and, in fact, most of the teams that are using our agents are hiring more people. So we’re seeing ROI. We’re seeing opportunities that publishers are identifying across the funnel, whether it’s to increase productivity or to create more outcomes on their content.

The ROI is going to get even more and more accentuated, become more and more tangible because of the innovation cycles that are happening with AI. AI is constantly becoming cheaper to run. AI is constantly becoming better.

PH: Do publishers need to start taking a slightly longer term view?

MM: I wouldn’t disagree with that, having a long term view is good, but with a short-term milestone.

If you don’t have a target – let’s say in three months, testing something to increase session time on page by 10% – unless you have those short term KPIs as well, the long term would just feel like a dream. You will not be able to get alignment, interaction and momentum within your teams.

Each media company has a long term – thinking that AI will be valuable within their tech stack, or within their analytics stack, or content stack, whatever that use case they are thinking. But having short-term goals is critical to be able to achieve the long term vision.

The maximum I always say for a project is to have a target for the next three months. Every quarter there should be a very tangible target. That doesn’t mean that every time it doesn’t get met you get rid of people. Tech teams really need to learn what went wrong.

I think everyone in the industry would agree, AI is going to change operations. It’s going to change how media companies monetise. It’s just a matter of choice. Do I want to play the waiting game and see who succeeds or fails? Or do I want to be the person that controls the narrative?

This Media Briefs episode is sponsored by Bridged Media, democratising AI for publishers. Through no-code AI solutions, Bridged lets publishers access the power of machine learning and Gen AI to meet their engagement and revenue objectives.

Publisher-first AI tools detect where the audience is most likely to engage and through a single line of code, introduce action cards that prompt readers to register, subscribe, or read more, helping publishers establish richer relationships.

Learn more about Bridged Media’s no-code AI tools on their website.

More good reads from the week: