Yes - since July 2025, and we've watched it happen inside real measurements. Here's what that means for your brand, and how much it's actually worth.
The split between engines is the story. Across the measurements we run, Instagram pages turn up as cited sources on a minority of answers, and the engines disagree sharply about it: some cite Instagram regularly, others essentially never, even though the same content is indexed and available to all of them. Brands' own profiles do get cited, unprompted.
So it is real, and it is uneven. That combination is exactly why it belongs in a plan as a supporting surface rather than a lead one - cheap to do properly, and never the thing you build a strategy on.
Treat your public profile as a retrievable surface, not a feed. Captions and alt text are editable any time - so your pinned posts should answer the questions buyers actually ask: where to try the product, delivery, price range, what the certifications mean. Question first, plain answer under it. One correction worth knowing: Stories and Highlights are not indexed - Instagram asks search engines to leave them alone - so keep Highlights tidy for the buyers who land on your profile, but the retrievable payload is captions, alt text and bio. It's an afternoon of editing what already exists.
"Does AI read X?" almost always has the same honest answer: partially, unevenly across engines, and less than the platform's press release implies. What matters isn't whether AI can read a surface - it's whether it does, on the questions your buyers ask. That's measurable, question by question. Which is exactly the measurement we run.