Is AI the Cure to Software We Hate?
Klarna is showing AI can free companies from bulky software platforms. For newsrooms, the benefits are different.
Amid all the excitement over the release of OpenAI o1 — an AI model that's said to be capable of better reasoning than any of the company's previous models — you'd be forgiven for not hearing about Klarna.
If you're like me, you only know what Klarna is because it's recently become the poster child for applying AI in a business. Every few months the company declares some new efficiency it's gained from deploying an AI tool, the most recent being its decision to kill contracts with two major software providers because Klarna found its AI systems could do the job just as well if not better. Could this be a potential template for everybody, including the media?
More on that in today's newsletter, but before I get to that, a quick reminder: I'll be going deep on how AI is altering the way we get information at a happy hour fireside chat for Hotwire Global in New York City on Wednesday. If you're in the City and are interested in learning more about the role AI-powered search engines like Perplexity and SearchGPT will play in media and communications, get in touch.
Also, I'm excited to attend the ONA conference in Atlanta this Thursday and Friday. I hope to meet and learn as much as I can from fellow journalists and content creators who are immersed in how AI is being applied to newsrooms, journalism, and content creation more broadly. If you’re going to be there, I’d love to meet up.
Finally, The Media Copilot's upcoming AI training classes for media professionals are sneaking up! Whether you're just starting your AI journey or want to level up your game, I guarantee there's a lot you'll get out of it. AI Quick Start, our 1-hour introductory course, is happening Sept. 25, and our popular AI Fundamentals class — three hours of intensive training on prompts and tools — is on Oct. 4 (use discount code AIFALL for 50% off). Or if you'd prefer a custom training session for your team, go ahead and shoot us an email.
AI Scams Are Rising. Here’s How You Can Protect Yourself.
Scammers just got even more dangerous thanks to AI. It’s become incredibly easy to copy someone’s voice or create a deepfake. Many cases have already been reported of people impersonating family members asking for money.
Here’s how you can help prevent that: Incogni is a personal data removal service that scrubs your personal information from the web. Incogni:
Protects you from identity theft and scammers taking out loans in your name.
Prevents strangers from buying your personal information on search sites.
Get 55% off with the code COPILOT. And if you’re not happy, get a full refund within 30 days.
AI vs. Enterprise Software
If you're looking for evidence that AI is making workplaces more efficient, Klarna would like to talk to you. The Stockholm-based payment provider has been making a name for itself with its hard push into generative AI. A couple of weeks ago CEO Sebastian Siemiatkowski said AI is enabling the company to "do more with less," and that he'd like to cut the company's workforce in half as a result. (He clarified that no layoffs were planned, but that the company would not replace workers who left.)
Now the company is stealing headlines again — this time for publicly announcing it would ditch Salesforce and Workday, two large enterprise software providers, since Klarna discovered it could replicate what those platforms do with AI tools.
Some reacted with skepticism, including Marc Benioff, CEO of Salesforce. You probably saw that coming, but Benioff has a point: Turning away from popular software platforms means you don't benefit from new features, upgrades, and standardized training. Plus employees experienced with that software suddenly have useless skills. And while big deployments of those platforms can get expensive, you probably won't save much since now you'll need to employ people to build and maintain those in-house systems.
Others made the case that moving away from bloated, heavy software platforms is exactly the kind of efficiency AI creates. What Klarna has done is create their own custom systems so their employees can (presumably) more easily use their private data. By doing so, Klarna can tailor its systems to focus only on the features they actually use and save on subscription fees in the process.
Democratic Processing
To me, this is the democratizing effect of AI in action, but applied at the organizational level. My friend Harry McCracken at Fast Company has written about the idea of self-aware software: the idea that a feature like Generative Fill in Photoshop lets a user who may not be an expert in the application perform tasks above their skill level. Get good enough at prompting — which is to say, talking to the AI — and you might even do better than someone with midlevel skills, and certainly much faster.
Broadly, the point is that someone with limited technical ability but is very familiar with the nuances of the AI system will often get better output than a highly technical person who isn't.
That's what I think is going on with Klarna, but at the organizational level. As a company, Klarna knows the metrics that drive its business, and it needs systems that help drive those up and to the right. Software is simply a means to that end. There are risks with taking a path with generative AI — with concerns of privacy, hallucinations, and bias — but if you can account for those, there's a good chance you'll get to your goal faster. And, since GenAI is all about natural language, with more people along for the ride.
Let's not get too excited. Like any system, generative AI depends on good data, and even better data management. If your organization doesn't have that sorted, the output from AI systems would likely be poor — or worse.
Can the Media Have Klarna-like Goals?
Also, for the media specifically, several publications have tried to reshape themselves, at least in part, as tech companies over the last decade. It usually doesn't go well. So I'm by no means suggesting any publisher should start thinking they can replace their CMSes and video players with ChatGPT.
However, the democratizing effects of AI still apply to media, just at an informational level. Instead of software, journalists can bring the power of natural language to get what they need from a set of information.
Finding the best information for your story — the right data point, the right source, the right nugget within a document — has traditionally been one of the most time-consuming parts of journalism. But with access to the right prompts and the right tools, that hard work of pointing yourself in the right direction could be done a lot faster than before.
Take court reporting: Finding the right documents and then translating what happened in a legal hearing into the language readers will understand requires proficiency in very specific public data platforms (PACER, etc.) and a lot of time reading. While AI can't yet match the analytical skill of a seasoned journalist, it does reduce the barriers to getting the right information quickly.
Having the right tools is an important component. You'll only get so far with Claude or ChatGPT. Using AI effectively with your own data will almost always mean using a tool that's trained, or has access to, specific data, with highly tailored prompting. Chat might be a part of the interaction, but not the only, or even the most important one.
Although "getting leaner" is often a euphemism for layoffs, it's an apt analogy when thinking about using AI in this way. Efficiencies aren't just about cutting things out of the process, like with Klarna. Getting and staying lean is just as much about what you're eating, or putting into the system. With the right data and tools, journalists can use AI to quickly surface the right information. That in turn feeds the best story-creation platform available: themselves.