AI Search

When AI Answers the Coupon Question: What It Means for Brands

November 2026·6 min read

A user types a retailer's name followed by 'discount code' or 'promo code' into Google, ChatGPT, Perplexity or Gemini. The AI engine surfaces a code directly in the answer. The user copies it and moves on. No visit to a coupon aggregator. No click to your website. No session recorded in GA4.

This is already happening. Google AI Overviews, ChatGPT and other AI engines are answering promotional code queries inline, pulling codes into the response itself. It is a small but telling example of AI search consuming a category of query that previously sent traffic somewhere - often to Vouchercodes, RetailMeNot or similar intermediaries.

Who Loses When AI Owns the Coupon Query

The obvious losers are the coupon aggregator sites. If AI engines are pulling codes into their own answers, the affiliate click that previously drove revenue for those platforms evaporates. That is their problem to solve. But brands are not exempt from the consequences.

If a user retrieves your discount code from a ChatGPT response, applies it at checkout, and completes the purchase, that order lands in your CRM with no meaningful source attached. Your GA4 session might show direct, or nothing at all. The attribution chain is broken before it starts. You cannot credit AI search with a conversion you cannot see.

There is also a subtler issue: if AI engines are sourcing outdated or incorrect codes - expired promotions, codes meant for specific customer segments, or codes from third-party sites you never authorised - your brand is implicated in a poor customer experience you did not orchestrate.

What AI Engines Are Actually Extracting

AI engines do not conjure promotional codes. They extract them from somewhere - your own site, affiliate publisher content, deal aggregators, press releases, social posts, or any publicly accessible source that has indexed the code. The question for any AEO programme is: are you the authoritative source for your own promotions in AI search, or are third parties beating you to it?

If you publish promotional offers on your own website - a dedicated offers page, a seasonal landing page, a clearly structured promotions section - and that content is crawlable, well-structured and updated, there is a reasonable case that AI engines will cite your site over an aggregator. Structured content with clear entity signals (brand name, offer terms, expiry information) is more extractable than a loosely formatted third-party listing.

Schema markup plays a role here too. The SpecialAnnouncement and Offer schema types within schema.org allow you to mark up promotional content with machine-readable fields: price, availability, valid-through dates. Whether AI engines are consistently reading and prioritising schema-marked promotional content remains a matter of ongoing observation, but providing the signal costs nothing and may improve extraction accuracy.

Structured Promotional Content as an AEO Asset

Most brand websites treat promotions as campaign ephemera - banners that go up during peak periods and disappear when the sale ends. That approach works for human visitors who see the banner in context. It is poor practice for AI extraction, which depends on consistently available, well-labelled content rather than time-limited page elements.

A permanent, well-maintained offers or promotions page - with clear headings, human-readable offer descriptions, defined validity periods and schema markup - gives AI engines something stable to cite. When the AI is asked about your brand's current discounts, you want it pulling from your page, not from an affiliate publisher who last updated their listing six months ago.

This is the same principle that applies across AEO: being the most structured, most accessible, most authoritative source for information about your own brand. Promotional content is simply another category where that principle applies. The brands that treat their offers page as a genuine information asset - kept current, marked up properly, internally linked from product pages - are better placed to earn the citation than those whose promotional content only exists in campaign banners.

The Attribution Problem Needs a Practical Fix

If AI engines are serving coupon codes inline and users are applying them at checkout, you will see the effect in redemption data before you see it in traffic data. Promotional code redemptions that cannot be tied to a marketing source - no referring URL, no UTM, no tracked session - are worth auditing. A spike in direct-attributed or unattributed conversions during a period when you are running a promotion is a reasonable signal that AI search is involved.

The more practical fix is code segmentation. If you can issue distinct promotional codes for different channels - including a code variant that only appears on your own website's offers page - you can identify when that specific code is redeemed without needing referral tracking to do it. The CRM records the code; you know the source. This is a low-tech solution that works independently of whatever AI engines choose to do with their referral headers.

For teams already running structured AI attribution work - monitoring GA4 for referrals from known AI engine domains, tagging AI-sourced sessions, mapping them through to CRM outcomes - the coupon scenario is a reminder that not all AI-influenced conversions leave a session fingerprint. Revenue attribution in AI search requires both traffic-side tracking and conversion-side code or identifier tracking.

What This Signals About Transactional AI Queries

Coupon queries are transactional. The user is mid-purchase. They are not researching; they are completing a transaction and looking for a marginal advantage before they do. The fact that AI engines are now handling these queries inline - rather than routing users to specialist sites - tells you something about where AI search is heading on the intent spectrum.

Earlier observations in this space focused on AI search dominating informational and definitional queries. Transactional queries were considered safer ground for traditional organic and paid results. The coupon example complicates that view. If AI engines are absorbing price-sensitive, action-oriented queries, the category of query considered immune to AI answer displacement is narrowing.

For AEO strategy, the implication is that content structuring efforts should not be limited to guides, FAQs and explanatory articles. Promotional pages, product specification pages, pricing pages and other transactional content assets all deserve the same treatment: clear structure, accurate entity signals, appropriate schema, bot-accessible crawling. Wherever users might ask AI search a question that your website is best placed to answer, your content needs to be extractable.

The Practical Checklist

If you are running an AEO programme and you have active promotions in market, run through these basics. First, do you have a dedicated, human-readable offers page that is indexed, crawlable and kept current? Second, is your promotional content marked up with Offer or SpecialAnnouncement schema where appropriate? Third, are your promotional codes segmented in a way that allows CRM-side attribution independent of session tracking? Fourth, have you audited what AI engines currently surface when a user asks about your brand's discounts - and is the information accurate?

None of this requires new technology. It requires treating your own promotional content with the same structural rigour you would apply to any other content asset you want AI search to cite accurately. The brands that do it now will own the citation. The ones that do not will find their promotions cited from sources they did not author and cannot control.