TLDR;
Google announced Direct Offers, a pilot that shows eligible advertiser offers in AI Mode when shoppers are close to buying. It gives retailers another way to encourage a purchase. The important test is whether it generates extra profit after discounts and advertising costs.
What happened
Google announced Direct Offers, a pilot for eligible advertiser offers in AI Mode when users are close to buying. It adds a commercial incentive within conversational shopping, with advertiser eligibility and later expansion distinguished from broad availability.
Why it matters
A conversational incentive appears near a customer's purchase decision, which makes it commercially interesting and difficult to evaluate through conversion rate alone. High-intent users may convert without the offer. Enterprise brands need to account for discount cost, eligibility, acquisition quality and the reporting limits of the pilot.
A tailored offer inside AI Mode can reach a shopper while a buying decision is still being formed. That timing may be valuable, particularly if the offer resolves a genuine obstacle such as total cost. It can also subsidise a sale the brand would have won anyway. Enterprise retailers need to distinguish persuasive relevance from profitable incrementality. A discount that produces more orders is not automatically a good campaign if it reduces contribution margin or teaches existing customers to wait for an incentive. The pilot status matters because access and reporting may be limited. The sensible judgement is to test a defined commercial hypothesis with the available evidence, not move a large promotional budget on the strength of a new format. Offer design should follow category economics and customer needs.
How your brand can benefit / be affected
Define eligible products, discount limits and stock safeguards before participating. Keep offer conditions consistent with the landing page and checkout, including expiry and customer restrictions.
Compare contribution after discount and media costs with an appropriate baseline. Review new-customer value, cancellations and cannibalisation where the available data permits. Do not extrapolate pilot performance across products or markets that have not been tested.
Choose products or customer situations where an incentive has a defensible purpose. Set an acceptable margin after the discount, fulfilment costs and likely returns, then decide which offer could improve the buyer's decision without exceeding that limit. Ensure eligibility, expiry and exclusions are clear wherever the supported experience presents the offer. If the promotion applies to only some variants or requires a minimum purchase, the destination must make that obvious before payment. Coordinate pricing, merchandising and paid media so the platform offer does not conflict with the merchant's own promotions or create a confusing promise that customer service must resolve.
Agree the test's decision rule before launch. Compare profit, new-customer contribution and purchase quality with a credible baseline or comparison where the available setup permits one. Review whether discount use is concentrated among customers who were already likely to buy. Keep reporting limitations explicit rather than assigning certainty to an incomplete view of the journey. If the pilot cannot answer the incrementality question, limit spending and use it to learn about execution and demand. Do not interpret redemption alone as success. The brand should expand only when the offer produces a better commercial outcome than the alternative use of the same promotional budget.
News date: 11 January 2026. Editorial review: 16 September 2026. Analysis includes subsequent developments where stated.