What are conversational AI ads? Preparing for Agent i
Conversational AI ads are advertisements shown in relation to the interests and conditions developed during an AI conversation. This guide uses Japan's Agent i to explain eligibility and the information advertisers can prepare.
Advertisers cannot select Agent i as a dedicated destination and guarantee delivery. Start by checking eligibility for the supported ad products, then align the advertisement with the purchase or inquiry page.
Eligible ads and delivery conditions
On September 1, 2026, LY Corporation announced expanded advertising related to Agent i conversations. Search-linked Shopping Ads expanded with advertiser billing from August 31; auction-based Display Ads were serving a subset of users, with billing from July 23. Official announcement
Eligible advertisers can participate through those Shopping or Display ad products. However, the announcement says Agent i cannot be individually selected as a destination. Ads may appear depending on delivery conditions.
Prepare in this order:
- Confirm the appropriate ad product and its eligibility requirements.
- Reconcile the advertisement, product information, and landing page.
- Check the campaign's budget and outcome measurement.
- Establish which placements and reporting scopes you can identify.
Use current account settings and official guidance without assuming dedicated bidding controls or a dedicated report exist.
Distinguish paid advertising from AI-answer references
Agent i's help describes ads as independent of its answers, with labels or frames identifying them. Users' questions and answers are not shared with advertisers. Official advertising help
An ad appearing and an AI answer referencing your business are different observations.
| Touchpoint | What to prepare | What to inspect |
|---|---|---|
| Paid advertising | Eligibility, ads or product information, destination, budget | Reported delivery, clicks, and outcomes |
| AI-answer references | Accurate public information and useful explanations | Observable references and visits |
| Purchases or inquiries on your website | Terms, application process, measurement | Purchases, qualified inquiries, cancellations |
More ad clicks do not prove more recommendations in AI answers. For product information, see the agentic-commerce product-data checklist.
Connect buying questions to the advertisement and landing page
Use real questions your business receives and the conditions buyers compare. The following editorial worksheet checks your content; it does not reproduce Agent i's delivery rules.
| Example buying question | Check in the ad | Check on the landing page |
|---|---|---|
| Can I use this where I live? | Accurate service-area description | Area, exclusions, contact route |
| Does it fit my intended use? | Clear use case and intended buyer | Specifications, compatibility, comparison information |
| Will it arrive in time? | Evidence for delivery promises | Stock and fulfillment conditions |
| What is the total cost? | Understandable price conditions | Extra charges, shipping, contract terms |
| What happens after I apply? | Clear next action | Purchase, booking, or inquiry steps |
Start with one product and check whether the ad and page give consistent answers. Review existing promises against actual sales conditions before adding separate AI-oriented text.
If you use AI for this work, provide verified product facts and anonymized questions your business holds. Do not assume access to users' Agent i conversations.
Help prepare an advertisement and landing page that answer buying questions.
Product or service:
Verified price, area, delivery time, and eligibility:
Actual customer questions, with personal information removed:
Current ad copy and product information:
Landing-page text:
Return:
- Buying questions and answers supported by the inputs
- Conflicting explanations between ad and page
- Conditions needing confirmation and questions for the owner
- Suggested revisions using verified information only
Do not invent Agent i conversations or delivery results.
Do not promise an AI recommendation or guaranteed ad placement.
Connect observable delivery to business outcomes
Before evaluating performance, check whether your ad product lets you distinguish delivery on Agent i. An aggregate campaign report that cannot separate placements is not a report of Agent i's performance alone.
Keep these records:
- Platform: dates, spend, impressions, clicks, and available placement breakdowns.
- Website: landing pages, identifiable advertising traffic, applications, and purchases.
- Business: qualified inquiries, closed sales, and cancellations.
- Changes: revisions to prices, ads, product information, pages, and budgets.
If both spend and inquiries rise, a new placement is not necessarily the sole cause. Leave inseparable effects unobserved and evaluate campaign-wide costs and outcomes first. The offline-conversion introduction explains linking later sales to advertising; its implementation steps apply to Google Ads, so use the appropriate LY Ads specifications for implementation here.
Your first task is to confirm ad eligibility and compare one product's advertisement and landing page. Shoryaku AI's chat can help discuss messaging and outcome interpretation using the aggregate figures and business information you provide. LY Ads is not automatically connected; prepare your own summaries.