Ads in AI apps.
Surfacedd’s network spans third-party AI apps across four modalities. Here’s how to run there.
The network, described.
Surfacedd aggregates third-party AI apps into a single ad network. On the publisher side that means one SDK, one disclosure standard, one revenue share, one ads policy. On the advertiser side it means one account, one creative pipeline, one auction, one reporting dashboard across every Surface served in any app in the network. The integration work that would otherwise happen app by app, contract by contract, disappears into one layer.
The apps that run Surfacedd cover the four modalities the AI layer is organized around. Text-first chatbots and writing tools carry the largest inventory pool. Image generators and design tools carry the next-largest. Voice-first assistants across in-home, in-car, and phone contexts carry a narrower but growing pool. Code completion and IDE-integrated coding tools carry the smallest pool and the highest-intent audience. New apps join the network on a rolling basis and appear in the advertiser console when they clear technical and disclosure review.
The network thesis is that the AI layer fragments rather than consolidates. If every specialized workflow ends up in a dedicated app, the only durable reach shape is a network that spans them at the Surface level. Advertisers who buy one app at a time do not keep up as the category splits. Advertisers who buy the network do. For the broader audience argument, see reach AI users.
The four formats and where each fits.
Four Surface types, one creative pipeline each. Pick the formats that fit your creative and your goals. Brands running across all four get the widest artifact coverage; brands running one or two get the cleanest creative fit.
Text in chatbots and writing tools.
Text Surfaces live inside chatbot replies and writing assistant outputs as labeled sponsor units beside the organic answer. Direct response, branded discovery, evergreen product messaging, and local services all fit. The format is the widest inventory pool and the lowest spend floor on the network. Most advertisers start here and expand into the other formats once the first campaigns clear. See AI brand placement for the placement detail.
Image in generators and design tools.
Image Surfaces live inside AI-generated images. A user prompts a scene; the generator composites a named product into it with a visible sponsor credit. Consumer goods, fashion, home, and travel brands fit the format because their creative already exists in product-in-scene form. The unit does not overlay on top of the composition; the sponsor credit is inside the frame with a consistent placement the app publishes.
Voice in in-home, in-car, and phone assistants.
Voice Surfaces are 5 to 10 second disclosed spoken segments, bracketed by explicit cues before and after. Branded audio, short calls to action, and local businesses fit the format. The assistant does not fold the sponsor into its organic answer; the segment is a separate piece of audio with its own labeling. Production support is available for advertisers without existing audio assets.
Code in IDEs and completion tools.
Code Surfaces are labeled sponsor lines in the comment layer of a code completion. Developer tools, cloud services, APIs, observability, and infrastructure brands fit the format because the audience writing the code is the audience buying the product. The sponsor never goes inside executable code. Code inventory is the smallest pool on the network and the highest price per impression.
Targeting without tracking.
Targeting on Surfacedd is contextual. Relevance is decided at serve time by the prompt the agent is handling and the Surface the output ships on. The prompt is the signal; the surface context is the filter. The user is not a subject of the transaction; the Surface is.
Advertisers choose context buckets, app inclusion and exclusion lists, geography, and pacing. Context buckets include conversation topic, user-expressed intent, product category, and workflow stage. Buckets are published in the advertiser console and can be combined. An electronics retailer running image Surfaces can target kitchen-appliance prompts in cooking apps and exclude restaurant-equipment prompts in professional tools; both filters run at serve time without user identifiers.
What the network does not do: third-party cookies, cross-site IDs, device graphs, fingerprinting workarounds, first-party to third-party identifier bridges, or identity resolution across apps. None of those mechanics are part of the pipeline. The targeting that exists works because the prompt is a rich, fresh signal that does not require a persistent identifier to be useful. The target is the context of the request, not the person making it.
This architecture survives regulatory changes because it never depended on the mechanics regulators are restricting. Cookie deprecation, mobile identifier rollbacks, browser privacy enforcement, and data protection law all leave the pipeline untouched because the pipeline does not collect the data in the first place.
Reporting and measurement.
Reporting is aggregate, by Surface, by app, and by context bucket. The dashboard surfaces impressions, clicks, CTR, CPC, CPM, completion rate for voice, and Share of Placement against the advertiser’s declared competitive set. Reporting refreshes hourly during live campaigns. CSV export and API access are standard. Historical series are retained for 24 months.
Share of Placement is the primary reach metric. The count of disclosed sponsor Surfaces a brand occupies across the network over a period, divided by the count of Surfaces available in the brand’s competitive set over the same period. See the Share of Placement page for the formula and the comparison to share of voice in traditional media.
What the network does not report: user-level click paths, per-user frequency, per-user conversion attribution, or cross-app behavioral graphs. The reporting stops at aggregate because the architecture stops at aggregate. Advertisers who need per-user attribution in other channels can still run those channels; Surfacedd reporting sits next to them rather than trying to replicate their model.
For campaigns running alongside first-party AI ad products, the methodology for combining signals is published and updated as the market matures. Reconciliation today is manual and transparent; reconciliation tomorrow will be closer to automatic as the shared measurement standards harden.
Pricing overview.
Pricing is published per Surface type. Ranges reflect the current floor. Clearing prices are set at auction at serve time.
- Text Surfaces: CPC from $0.50, CPM from $5. No campaign floor on self-serve.
- Image Surfaces: CPM from $8. Minimum campaign $2,500.
- Voice Surfaces: CPM from $10. Minimum campaign $5,000.
- Code Surfaces: CPM from $12. Minimum campaign $5,000.
- Dayparting, pacing caps, and geography are standard on every Surface.
- Inclusion and exclusion app lists are standard on every Surface.
CPM ranges reflect inventory scarcity. Code sits highest because the pool is narrow and the audience converts. Voice sits next because of creative production cost and inventory limits. Image sits after voice. Text sits lowest because the pool is deepest. Every Surface runs the same auction; the floor is the lower bound of the clearing range.
For context on how pricing fits alongside first-party AI ad products, see reach AI users. For publishers considering joining the network, see monetize your AI app for revenue share and integration detail.
Frequently asked questions.
What counts as an AI app on this network?
Can I pick which apps to run in?
How fast can I launch?
Is reporting app-level or aggregate only?
Do you support programmatic buying?
What happens if an app violates the disclosure rules?
Reach users inside the AI tools they already use.
CPC from $0.50. CPM from $5. Text, image, voice, and code placements across independent AI apps.