Koah

Koah raised $5 million in a seed round in September 2025, led by Forerunner Ventures with participation from South Park Commons and Andrew Karam, co-founder of AppLovin.

Before the seed round, Koah completed a pre-seed round backed by South Park Commons. The amount raised in the pre-seed round was not disclosed.

Koah develops an advertising network tailored for AI chat and agent interfaces, diverging from traditional web pages or mobile feeds. The platform operates as a lightweight SDK that AI app developers can integrate into their products with minimal code.

Developers register through Koah's dashboard, obtain a unique Publisher ID, and integrate one of the open-source SDKs available for JavaScript, React, React-Native, Flutter, iOS, and Android. For example, a complete Flutter implementation requires approximately 10 lines of code, where developers initialize the Koah instance during app startup and wrap LLM responses in a specified widget.

During runtime, the SDK transmits end-user queries and response metadata to Koah's ad server. The platform's ContextMatch model applies embeddings and first-party telemetry to evaluate queries for commercial intent. If intent is high, the server delivers a JSON ad object; if intent is low, it returns null, ensuring the app remains ad-free.

Impressions are recorded only after at least one second of on-screen visibility, verified through integrated viewability tracking. Ad formats are specifically designed for conversational interfaces and include prefix/suffix text ads placed before or after model responses, sponsored citations displayed in citation panels, follow-up questions aimed at increasing engagement, and static user-targeted placements in input screen white space.

Koah operates as a B2B2C advertising network using a revenue-sharing model with AI application publishers. The platform functions as an intermediary between advertisers targeting AI chat users and publishers monetizing their conversational AI applications.

Publishers integrate Koah's SDK and generate revenue when ads are displayed and clicked within their applications. Koah retains a percentage of advertising spend as its fee, consistent with standard ad network economics, where platforms typically keep 25-35% of gross advertising revenue.

Advertisers utilize a self-serve campaign builder that supports CPC, CPM, and affiliate CPA pricing models. The platform includes pixel-based conversion tracking compatible with manual implementation, Google Tag Manager, Webflow, and Framer. This setup enables advertisers to pass revenue and value parameters back to Koah, facilitating campaign optimization.

The business model leverages sub-second ad load times and higher engagement rates compared to traditional display advertising. These performance metrics allow Koah to achieve premium eCPMs while delivering improved outcomes for advertisers, contributing to a competitive advantage in the AI advertising market.

The platform's technical infrastructure is built on Ruby on Rails and TypeScript, with AWS backend services including RDS, S3, and Kafka, as well as ClickHouse for analytics processing.

Google has started integrating ads within AI Overviews and AI Mode, utilizing its demand-side tools, first-party intent data, and extensive advertiser network. The company places text, shopping, app, and local ads above, below, or within generative answers, potentially diverting incremental budgets from independent networks.

Amazon Ads is introducing generative audience-building tools and AI-powered video creative features to retain brands within its walled garden ecosystem. ByteDance operates AI chatbots like Doubao, which has over 60 million monthly active users in China, and could adapt its established advertising stack to compete directly.

OpenAI is evaluating in-house advertising models that, if implemented within ChatGPT, could scale rapidly and reduce the independent inventory pool aggregated by networks like Koah.

Nexad raised $6 million in seed funding from a16z and Prosus, describing itself as an AppLovin for AI applications. The company develops and places contextual ads using LLaMA, Gemini, and Anthropic models, reporting 30 million users across seven AI partners and proprietary ad-generation technology.

OpenAds provides web-based snippets that create unique ads per impression with privacy-first contextual matching and in-ad conversation features. The platform focuses on creative automation and compliance with cookieless privacy standards in response to shifting regulatory and consumer expectations.

Established platforms like Tavus have enhanced AI-driven video capabilities for marketing, sales, and communication, influencing personalized engagement strategies that may intersect with conversational advertising. These platforms leverage advanced AI recommendation systems and content personalization expertise, which could be adapted for chat-based advertising formats.

Koah can expand its offerings beyond its current suffix, prefix, sponsored-citation, follow-up-question, and static display units to include voice answers, AI-generated images, and agent actions. These additions would enable monetization of the growing multimodal and agentic layer of GenAI, increasing inventory without requiring additional publishers.

Integrating lower-funnel commerce tools such as built-in checkout links, affiliate SKU feeds, or shoppable carousels could capture transactional GMV alongside CPM and CPC revenue. This approach would broaden Koah's TAM from advertising spend to encompass AI-commerce enablement.

Developing AI-powered creative generation capabilities could mitigate competitive pressures while accessing long-tail SMB budgets that lack internal design resources.

The rapid proliferation of LLM wrappers, vertical assistants, and agent frameworks—thousands of which launch monthly—presents a significant growth opportunity. Offering drop-in SDKs for widely used frameworks like LangChain or Vercel AI SDK could increase the publisher base by 10-20x.

Enterprise and employee chat remains an underutilized market, with 58% of US adults under 30 reporting use of ChatGPT at work. Privacy-safe internal sponsorship units for enterprise chatbots could redirect B2B lead-generation budgets currently allocated to platforms like LinkedIn and Google Search.

AI applications outside the US prosumer market also represent a meaningful growth area, particularly in regions where users are less inclined to pay $20 monthly subscriptions for AI services.

While Koah already operates globally, there is potential for deeper expansion in Latin America, India, Indonesia, and Nigeria, which collectively account for over 1 billion English-capable mobile users. Localizing SDKs for low-cost on-device models and partnering with regional resellers could unlock these high-growth markets.

Expanding beyond tech-focused advertisers such as Upwork and Skillshare to include retail media networks or travel OTAs would increase effective CPMs in non-software categories and reduce reliance on any single advertiser vertical.

Platform dependence: Koah's business model depends on sustaining relationships with AI application publishers, who may opt to develop proprietary advertising solutions or collaborate with larger platforms such as Google or Amazon. These platforms provide broader advertiser demand and offer higher revenue shares, which could make them more attractive partners.

Regulatory constraints: Increasing global regulatory scrutiny of AI-generated content and advertising may introduce new requirements related to disclosure, privacy, and algorithmic transparency. Compliance with these regulations could necessitate substantial platform adjustments or restrict certain ad formats, potentially diminishing monetization efficiency.

Market consolidation: While the AI application ecosystem is currently fragmented, consolidation around a few dominant platforms could shrink the pool of available publishers. This shift may also strengthen the negotiating power of consolidated platforms, enabling them to secure more favorable revenue-sharing agreements or create competing advertising solutions.