The team at Wing is fortunate to meet often with top product leaders at leading SaaS and AI companies, and those conversations always leave us energized about what’s happening – and what’s coming. Below are some of the insights we’ve been hearing lately about the landscape of AI-driven products and services today.
The need for speed in the AI world
First things first: speed is king in the AI world. Everyone nods along when the topic of execution speed and product velocity come up. It's like a high-stakes race out there, with all the leading companies trying to outpace each other on everything from major launches to small but consistent product updates.
Product convergence and the AI melting pot
Here's something interesting: these companies are starting to step on each other's toes. Notion's adding search features, Perplexity's dipping its toes in enterprise waters, and OpenAI? Well, they're just everywhere. It's like watching a game of AI musical chairs, and everyone's scrambling for a seat. Many of those we spoke to see these players becoming increasingly competitive.
The challenge of staying ahead in a commoditized world
Something everyone’s noticing: basic AI features are becoming a dime a dozen. Even major AI companies are talking about needing to become a full-on product company. The secret sauce? Better user experience, niche use cases, lightning-fast product updates, and smart distribution. It's not just about having AI anymore; it's about making the user experience better than the rest.
B2B AI innovation is top-down
Customers are proactively asking for AI-prescribed, top-down mandates – asking, "What AI features do you have?" – because they now have an AI budget to spend. In reality, the actual ROI and usage patterns vary wildly: some teams are using it heavily while others barely touch it. AI is the cool new toy everyone wants to play with. Those tasked with monitoring ROI haven’t yet started looking too deeply, but give it 18 months, and we might see some raised eyebrows and increased scrutiny.
AI usage in product teams: Not quite there yet
Surprisingly, product managers aren't as AI-crazy as you'd think. They're mostly using it for the usual suspects – writing docs, launch posts, and getting a grip on what users want. But here's a cool idea someone tossed out: imagine an AI that really gets your ideal customer, like a virtual product guru you can bounce ideas off. That’s something we’d love to see!
The innovator’s dilemma for SaaS
Some SaaS products are having a bit of an existential crisis. They're torn between sticking to their usual user-driven workflows and going full-on AI agent mode. That, though, can clash with the preferences of a product’s existing champions who own those workflows, which could create an opening for a more aggressive competitor. One product leader we spoke to summed it up like this: they know they should build more agent-centric workflows, but they're tiptoeing around it so as not to spook current champions.
Model hot potato and more eval orders
Evaluations are growing in importance as more and more models and features are launched and increase in their sophistication. Some companies are looking to companies like Braintrust for this. More and more, teams are increasingly being asked to mature and benchmark different models, and swap models as they evolve and improve. They're not messing with fine-tuning much yet, because they’re worried that since base models are still improving quickly, a fine-tuned model will be worse than the new base model improvement. It’s about building and shipping, versus achieving that last 1-5 percent optimization.
