I would not rebuild my app because Apple showed AI tools at WWDC 2023. Keynotes sell devices. They do not sign your paycheck next quarter.
Apple's early AI story was careful: on-device emphasis, privacy language, features woven into apps people already use. Useful direction. Slow real-world impact for random startups.
What small businesses could ignore for a while
Most local shops, agencies, and niche SaaS teams did not need to panic-integrate Apple's AI stack on day one.
Why?
- adoption follows hardware cycles
- APIs mature in steps
- user habits do not change because developers got excited
If your customers are on older devices, fancy on-device models might not matter yet.
What was worth watching
I paid attention to the shape of the bet:
- summarization inside familiar apps
- writing help where people already type
- image and voice features tied to OS permissions
- privacy as a selling point against cloud-heavy competitors
That tells you how Apple wants users to think about AI: helpful, local, branded, controlled.
The mistake is platform tourism
Teams burn months building "for Apple intelligence" before release details, limits, and review rules are clear.
They ship a demo. Apple changes a policy. The feature dies.
I would rather improve the core job my product already does than chase a keynote noun.
If you build on Apple's stack later
When APIs fit your use case, ask:
- does on-device limit quality in ways users will notice?
- what fails offline?
- how do updates reach old phones?
- does this feature matter to paying customers or only to your launch post?
WWDC is a trailer. Your business needs the full movie.
I would watch Apple's direction, ship boring value now, and integrate when users on supported devices are common enough to matter.
