Botify

Botify acts in the customer's own estate rather than only reading it: PageWorkers deploys technical optimizations directly onto the customer's live site, SpeedWorkers operates at their CDN intercepting and serving bot requests, and AgenticCatalog auto-deploys enriched product feeds to ChatGPT and Google. Around that a unified data model joins first-party crawl and log data with search console, analytics, competitive and paid sources, and Botify Assist reaches into the customer's existing tools through an MCP connection to pull real-time data.

Sourcebotify.com agentic search use case and the AWS partner architecture postread 2026-09-01

Botify Assist runs multiple agents that carry a workflow from analysis and insight through content creation to deployment, with SmartContent, SmartIndex and SmartLink named as agentic workflows and PageWorkers executing the deployment step at scale. The motion routes across systems rather than sitting in one screen: crawl and log data feed the analysis, generated optimizations are pushed to the live site, and enriched feeds are deployed outward to shopping protocols. Changes are instant and reversible, which is documented rollback on the deployment path.

Sourcebotify.com agentic search use case and the AWS partner architecture postread 2026-09-01

The corpus is the customer's own and the ingestion is the product. Botify crawls the brand's entire website, ingests its server logs, and joins that with search console, analytics and third-party data into a unified data model, then grounds content generation and Agentic Feeds in that crawl intelligence and cached content rather than in generic context. Rendered page content cached by SpeedWorkers feeds the same store, so what the agents write from is the customer's own site as it actually renders.

Sourcebotify.com agentic search use case and platform documentationread 2026-09-01

Review before action is stated as a product property: Botify's generative AI article says all AI generated outputs are recommendations only and require human review before any action is taken, and the agentic search page shows generated meta content being reviewed, deployed and split tested in a few clicks, with PageWorkers changes reversible. Indexation pushes and pre-rendering run without a per-action gate.

Sourcesupport.botify.com/en/articles/14846526-about-generative-ai-in-botifyread 2026-10-01

SAML single sign on through Okta, ADFS, Azure AD or OneLogin is enforced for every user on the account on Enterprise plans, while SCIM provisioning is stated as not supported. SOC 2 certification is named in Botify's generative AI article, and the vendor's own 2021 review describes a SOC 2 Type 1 examination by Schellman with Type 2 as the next step. The current report type and date are not stated.

Sourcesupport.botify.com/en/articles/9108628-configuring-ssoread 2026-10-01

Measurement is the platform's origin and it is deep: more than a thousand performance metrics computed from crawl and log data, AI Visibility tracking brand presence across twelve AI bots and platforms including ChatGPT, Perplexity, Google AI Overviews and Gemini, log analysis showing what search and AI crawlers actually fetched, and an alert panel raising early warnings on risks. None of that traces an agent; nothing reconstructs which agent ran, what it changed and why, and no execution log or audit surface for agent activity is documented, so what exists is observability of the customer's search estate rather than of the agents acting on it.

Sourcebotify.com platform pages and AI Visibility product descriptionread 2026-09-01

No agent memory is documented: nothing describes an agent retaining context from one optimization cycle or content generation to the next, and no memory layer with a stated scope or lifetime appears on the vendor's pages. The unified data model persists crawl history, log data and visibility measurement over time, but that store is the knowledge the agents are grounded in rather than state an agent carries between runs.

Sourcebotify.com platform pagesread 2026-09-01

Delivered as cloud software with no deployment option in the customer's own environment: no self-hosted, on-premises, private-cloud or single-tenant install is documented, and no residency commitment naming regions is published. Edge delivery through SpeedWorkers is a serving architecture at the customer's CDN rather than a deployment of the platform, since rendering and caching happen on Botify's own servers.

Sourcebotify.com SpeedWorkers architecture descriptionread 2026-09-01

A named suite of agents and agentic workflows ships and is browsable across the vendor's product pages: Botify Assist, SmartContent, SmartIndex, SmartLink, the Content Agent, PageWorkers, SpeedWorkers and AgenticCatalog, each aimed at a named job from content generation to indexation, internal linking, autonomous optimization deployment, edge pre-rendering and product feed enrichment. Agentic Feeds adds protocol-specific packaging for OpenAI's Agentic Commerce Protocol and Google's Universal Commerce Protocol.

Sourcebotify.com platform and agentic search product pagesread 2026-09-01

Work starts without a person asking for it: crawling and log ingestion run continuously, an alert panel raises early warnings when visibility or crawl behavior degrades, seasonal and trend shifts trigger automated content updates and instant indexation requests, and enriched feeds auto-deploy to ChatGPT and Google on their own schedule.

SpeedWorkers responds to bot requests as they arrive at the CDN, which is an event trigger in the strictest sense. The agent reaches the team through alerts and reporting and reaches the outside world by writing to the live site and to shopping protocol feeds, though its channels are systems rather than conversation surfaces.

Sourcebotify.com agentic search use case, alert panel and Agentic Feeds announcementread 2026-09-01

No customer-facing model choice is documented. The agents generate content and optimizations without any documented model selector, provider list, bring-your-own-key option or routing policy, and the vendor does not disclose which underlying provider powers Botify Assist. The MCP endpoints on both sides are tool and data connections rather than model routing.

Sourcebotify.com agentic search use case and developers.botify.comread 2026-09-01

Botify publishes first-party developer documentation at developers.botify.com, describing an API whose purpose is to make the platform's data available, covering thousands of metrics computed from website crawling, server log ingestion and third-party integrations. The REST API sits at api.botify.com/v1 with an OpenAPI specification and a query language for interactive queries and bulk export jobs to S3, Redshift, Cloud Storage or BigQuery, and a hosted OAuth-protected MCP server is operated at mcp.botify.com so external agents can work with Botify data alongside the vendor's own agents.

Sourcedevelopers.botify.com API documentationread 2026-09-01

A change can be evaluated against a criterion before it is kept. Split testing is built into PageWorkers deployments and meta content changes, so an optimization runs against a control and the difference is measured, and the vendor documents tools to measure the direct impact of enriched feeds on AI visibility and product selection. Deployments are instant and reversible, so a losing variant can be withdrawn. The test runs on live traffic rather than ahead of it.

Sourcebotify.com agentic search page and Agentic Feeds announcementread 2026-09-01

Rendering and crawling at scale are the product's foundation and both are documented. SpeedWorkers operates at the CDN level, reading, rendering and caching the customer's pages continually on Botify's own servers, intercepting bot requests and returning fully rendered versions of pages whose content would otherwise sit behind JavaScript.

Alongside it SiteCrawler crawls the brand's entire website and ingests server logs. Executing JavaScript to produce a rendered page requires a browser engine, and the vendor operates it, a headless fetch and render wired into the platform. No agent driving a browser or desktop session as a user is documented; the engine renders pages rather than navigating or acting inside them.

Sourcebotify.com SpeedWorkers product and dynamic rendering documentationread 2026-09-01