Every CPG marketer wants to know the same thing: if I only have time to optimize for one AI search platform, which one matters most? The answer is more nuanced than the popular narrative suggests.
We've pulled data from hundreds of brand audits at Chedder, plus public usage data and our own attribution work, and the picture in 2026 looks like this.
The raw scale numbers
Active users, monthly, US, as of mid-2026:
- ChatGPT: ~250 million weekly globally, ~100 million US monthly with shopping-adjacent intent
- Google AI Overviews: ~120 million US monthly users (Google search users who see AI Overviews on commercial queries)
- Perplexity: ~80 million monthly globally, ~30 million US
- Brave Search AI: ~15 million monthly users
- DuckAssist: ~10 million monthly users
- Apple Intelligence (with Apple Search): ~50 million monthly Apple Intelligence users in the US, but the shopping-query share is harder to estimate
ChatGPT is the largest by a meaningful margin. But scale isn't the same as shopper-relevance.
Which platforms actually drive purchases
Scale matters less than intent quality and click-through behavior. Across the brands we audit, the platforms break down differently when you weight by purchase-attributable traffic:
Google AI Overviews drive the most CPG clicks. Because Google Overviews appear ON Google's normal search results page, the user often clicks through to a brand site from the Overview citation. The conversion rate is meaningful.
Perplexity drives second-most clicks per query. Perplexity's interface heavily features citation chips that users click. A brand mentioned in a Perplexity answer with a citation link gets traffic at a higher per-impression rate than nearly any other AI platform.
ChatGPT drives less direct click-through than its scale suggests. Most ChatGPT users get the answer and act on it without clicking through. The brand recommendation lands as a name they recognize and search for separately (often via Google) or remember for later.
Brave Search AI drives engaged but small traffic. Brave's user base is small but privacy-aware and tends to have strong purchase intent. CPG brands in supplements, wellness, and tech tend to overperform on Brave.
Apple Intelligence is the dark horse. Hard to measure because Apple obscures referrer data. But the iPhone user base is highly purchasing-relevant and Apple Intelligence usage is growing fast.
The "scale vs intent" tradeoff in plain terms
If you want maximum brand IMPRESSIONS, prioritize ChatGPT visibility. The most shoppers see your brand name in answers via ChatGPT.
If you want maximum brand CLICKS, prioritize Perplexity and Google AI Overviews. These platforms surface citations the user actually clicks.
If you want maximum CONVERSION-WEIGHTED traffic, prioritize Perplexity and Brave. Smaller user bases but more committed clickers.
Most CPG brands should optimize for all four. The infrastructure overlaps enough that there's no real "choose one" tradeoff — getting your schema right, your AI crawlers unblocked, and your category Reddit presence built helps every platform.
What's different about each platform's recommendation logic
Brief breakdown of what each engine actually rewards differently:
ChatGPT
- Heaviest weight on Wikipedia and long-term entity recognition
- Less sensitive to fresh content
- Strong preference for brands with multiple consistent sources of information
- Will sometimes name a brand without any visible citation (drawing on training data)
Perplexity
- Heaviest weight on fresh content (last 12-18 months)
- Citation density matters more than entity recognition
- Strong preference for brands with recent press coverage
- Will name brands only if it can cite a source
Brave Search AI
- Privacy-aware index that excludes some surveillance-heavy sources
- Tends to weight independent review sites over major publishers
- Strong on niche category coverage, weaker on mainstream
- Often surfaces brands that don't appear in Google's mainstream results
Google AI Overviews
- Heavy weight on Google's existing SEO signals (PageRank, traditional rank factors)
- Brands that rank well in normal Google search tend to appear in Overviews too
- Strongest preference for "best of" listicle sources
- Most sensitive to schema/structured data of any platform
The optimization order for a CPG brand starting from zero
If you're starting from no AI search visibility and you can only fix one thing at a time:
- Robots.txt + AI crawler access (covers all four platforms at once; biggest single move)
- Product + Organization schema (most impactful for Google AI Overviews; helpful for all)
- Wikipedia + Wikidata entry (most impactful for ChatGPT and Apple Intelligence)
- Reddit category presence (impactful for ChatGPT and Perplexity)
- Fresh press cadence (most impactful for Perplexity and Brave)
- Independent reviewer relationships (most impactful for Brave; secondary for all)
The first three are technical. They can ship in a quarter. The last three are programs that need 6-12 months of consistent effort.
What about voice assistants and smart speakers?
Worth a note: Alexa, Google Assistant, and Siri are increasingly LLM-backed and increasingly used for product recommendations. The optimization is largely the same as for the platforms above — the engines under the hood are similar — but the user behavior is different (much more "just buy it for me" intent and less browsing).
Brands that win voice tend to be the brands that ALSO win on subscription-heavy categories (CPG essentials like detergent, pet food, coffee) where re-purchase is the norm. The first AI-recommended brand often becomes the default repeat purchase.
How to measure where you stand on each platform
Manually: open each platform and run 5-10 prompts about your category. Note which brands get named in which platform, and whether you appear. Repeat monthly.
Faster: run an audit at Chedder. We run real prompts across ChatGPT, Perplexity, Brave, and Google AI Overviews simultaneously, then report brand mentions per platform with verbatim excerpts. Takes under a minute.
The brands that dominate AI traffic in 2026 are the brands that treat each engine as its own channel, not as a single "AI" bucket. They have a different score on each platform, watch them independently, and prioritize fixes based on which engine is underperforming relative to where they should be.
Start with the audit. Pick the worst-performing platform. Fix the highest-leverage gap. Repeat monthly. That's the playbook.