No, It’s Not Just You: AI Product and Pricing Models Are a Hot Mess (And They Don’t Care Because They Have Us by the Balls)
Every week, a new AI model launches with groundbreaking features, confusing restrictions, and baffling pricing. Google and OpenAI are naming research tools the same thing, Microsoft’s Copilot pricing feels like 2005-era texting, and Perplexity’s latest upgrade comes with a major security risk. Here’s what’s happening—and what you need to know right now.
If you’ve been trying to keep up with AI tools for marketing and communications, you’re probably blown away by their capabilities right now—but also a little frustrated. And you’re not alone.
AI model offerings are a constantly shifting maze of new releases, restricted access, and baffling pricing structures. Consider this: Gemini’s Deep Research tool is powerful, but if your company has a Google Workspace account, you can’t use it yet. And when you’ll get access? Anyone’s guess. In the meantime, you either have massive FOMO or don’t know it exists. Desperate to check it out? You can open a personal Pro account, but the Deep Research functionality you just paid for might disappear a month later with the next model rollout. (Yes, that just happened.)
OpenAI isn’t much better. They just introduced what they are calling Deep Research. (No, you’re not crazy—OpenAI really did name this feature the exact same thing as Gemini’s research tool.) But, it’s only available to Pro users, not Teams users. (Not Microsoft Teams, ChatGPT Teams. I know.) ChatGPT Teams is meant for companies, while Pro is for individuals. If you run a business and want to provision a handful of employees to test the new feature, too bad. Each employee will need to create separate personal accounts just to try it out. The CISO will love it.
Then there’s the endless feature confusion. Can you attach documents? Maybe. Can it read a PDF better than a Word doc? Depends. Can it parse data from graphics? Good luck finding a consistent answer. And don’t forget, many major online sources of information block AI tools from scraping them altogether. So no matter what AI-based research you do, you still have to integrate manual research from those missing sources.
Perplexity’s pricing and feature mix is much clearer than many of its competitors, making it a solid choice for AI-powered research until now. But they recently added DeepSeek’s R1 model to enhance Perplexity’s capabilities. The results are beyond impressive. Watching the R1 model “think” through the research process is nothing short of incredible. However, the new reliance on R1 raises serious security concerns, because DeepSeek servers are located in the People’s Republic of China, and their privacy policy explicitly states that they are scraping all your data. So when using Perplexity—not to mention DeepSeek—you have to assume privacy is totally compromised.
Microsoft has a real opportunity to cut through the mess with Copilot, because Office 365 is already embedded in so many organizations. But first, it needs to fix the wildly impractical usage-based pricing just rolled out for Copilot features. Remember when cell carriers charged per text message, and your parents lost it when the bill arrived? That’s what Microsoft’s new Copilot pricing looks like today—except the bills will hit business units at the end of the month, not just teenagers with flip phones.
This chaos exists because AI models offer tools that are so valuable, early adopters will tolerate almost anything just to get access. The pace of innovation is outstripping these companies’ ability, willingness or need to focus on the usual go-to-market essentials, such as rational pricing models, clear feature sets, and naming structures that make sense. Let’s face it: when you have a tool this revolutionary, it sells itself. But long term? No matter how groundbreaking the technology, these companies should clarify their offerings and pricing.
So, what should you do? First, take a breath. Yes, AI pricing and feature sets are a confusing mess. Yes, it feels like these companies are making it up as they go. That’s because they are. But that doesn’t mean you should check out. We’re in sandbox mode, and while things are chaotic, these tools offer tremendous value you can leverage right now.
This is what it looks like when a massively transformative technology is rolling out in real time. The chaos is annoying, but it shouldn’t stop you from continuously learning, because these tools offer a dominant advantage to those who adopt them. AI isn’t just another software update – it’s a fundamental shift. So continue investing in learning and experimenting. Whether communications leaders like it or not, this is where the future is being built.