Filter one: does it change where my customer decides?
A trend matters to you only if it moves the moment or place your customer chooses a provider. Connected TV mattered because it moved television into a targetable, small-budget purchase. Most platform features do not move anything.
If the buying moment is unchanged, the trend is industry news rather than a business decision.
Filter two: can I test it for 15 percent and six weeks?
A trend that requires a rebuild, a new vendor, a content operation, or a six-month commitment before producing a read is not testable at local scale. That does not make it wrong — it makes it not yours yet.
Anything you can run for six weeks with a small share of budget and a clear cost-per-outcome read is worth trying if it passes filter one.
Filter three: is the case study a business like mine?
Impressive results from a national brand with a hundred-person team tell you nothing about a nine-person business in one county. The mechanisms that made it work usually do not exist at your scale.
Ask specifically for a case in your category, at your budget, in a comparable market. The absence of one is informative.
What passes right now, and what does not
Passes: county-level connected TV, programmatic out-of-home, streaming audio, and activating your own customer data. All testable, all cheap, all moving where decisions happen.
Does not pass for most local businesses: retail media without product, the metaverse in any form, NFT-adjacent anything, influencer programs below a certain scale, and "AI optimization" services promising placement in answer engines.
The honest position on AI: it has made execution dramatically cheaper, which is a real change to your costs. It has not created a channel you need to buy.