The onset of open source left many exploring whether to purchase ready-made AI technology off the shelf or instead leverage publicly available AI models to build their own capabilities. So we put the question out to some of the newsrooms we work with in order to establish an industry-wide view.
And their response? In short, there’s no one-size-fits-all solution. Intriguingly, when asked whether they’re more likely to build or buy generative AI capabilities, the response was a 50/50 split - highlighting the lack of a clear benchmark or direction within the newsroom community.
The truth is choosing between building AI capabilities in-house or purchasing a vendor’s ready-made solution depends on numerous factors like your specific needs, budget, technical resources, or goals as an organization. And it’s important for news media brands to fully understand their requirements before committing either way. So let's explore some key considerations for each approach that can help your organization come to the right conclusion.
As part of our study, we dug into the reasons why newsrooms opt for one approach over the other. And for those who told us they prefer to build generative AI capabilities in-house, a couple of clear reasons emerged.
Newsrooms that choose to build their own capabilities unanimously point to being able to control what they create (100%) - for example, when handling data. They also overwhelmingly point to how they can customize their solution (69%) as a key advantage.
No doubt building your own solution means you can tailor it to your specific data, terminology, and use cases. Organizations often see their custom-built models, algorithms, and training data as intellectual property, which can be a source of long-term competitive advantage.
Taken together, these data points suggest highly bespoke environments for these particular newsrooms. And that makes complete sense. If you’re a newsroom operating with especially nuanced needs, an out-of-the-box solution likely won’t fit the bill. And that’s where your in-house experts can get creative and create a tool that’s fit for purpose.
When we dug into why newsrooms would prefer to buy AI solutions off the shelf, once again the rationale was clear.
The most prevalent consideration was being able to leverage a vendor's position at the forefront of technological advancements (43%) as well as their likely ability to move faster than an in-house team (36%). This makes sense as there's no need to spend months on development, training models, and building infrastructure. With an out-of-the-box solution, you’re likely up and running in hours.
They also cited a lack of both internal expertise and technical resources (both 36%) as reasons to buy AI ready-made. What’s clear is that there’s a significant level of investment and effort required when building AI tools in-house - from the initial development effort to scaling up when required, as well as maintenance, updates and bug fixes which can consume a lot of time and effort.
The data reveals that, for those newsrooms that cannot maintain the upkeep, leaning on the ready-made capabilities of a vendor - whose dedicated teams of engineers and IT support are constantly improving and patching their models - makes much more sense.
In an interesting turn of events, both builders (38%) and buyers (36%) referenced cost savings as an advantage of their approach. The difference is likely in how and when they realize those savings.
For those who build, the initial investment is high. However, they might discover over time that the cost of their custom-built solution is lower, especially when usage volumes are high. For example, as they won’t need to pay per-minute fees to a vendor.
Whereas when buying solutions off the shelf, the upfront cost of a subscription to a vendor's service is much lower than the investment required for a custom-built solution, where costs include: in-house development teams, salaries, hardware, and ongoing maintenance.
Ultimately, the decision to build or buy generative AI is a strategic one that should align with your business goals and available resources.
Build your own capabilities if you: have specific and unique requirements; value control over data and customization; and have the financial and technical resources to invest in a likely long-term project.
Buy ready-made if you prioritize speed, ease of use, and access to state-of-the-art technology - and if development and maintenance will place a heavy burden on internal resources.
By carefully evaluating your organization's specific needs and weighing the pros and cons of each approach, you can make an informed decision that will provide the most value for your organization.