Measurement has always been the foundation of our work at Perrill. It's how we build strategies, make decisions, and prove results. We don't operate on guesswork, after all. As AI [search](https://www.perrill.com/google-shouldnt-reveal-search-algorithm/) started reshaping how people find information online, we knew we could help our clients show up there too. Recommending [generative engine optimization as a service](https://www.perrill.com/generative-engine-optimization-services/) was a strategic next step, and we knew we'd need an AI visibility measurement tool that gave us accurate data to inform our marketing decisions. We wanted the ability to explore AI search insights throughout the entire funnel, from prompt and mention reporting all the way to ROI. We saw the opportunity to build our own solution in-house—one that would support our reporting needs, give us full ownership of metric calculations, and leave room for continuous iteration as the AI search space evolves. Our [full team of in-house developers](https://www.perrill.com/what-we-do/web-developer-minneapolis/) has extensive experience building complex, custom software, so we got to work developing a platform to help drive better results for our GEO clients.  ## Introducing PAVE: Answer Engine Optimization PAVE is our proprietary AI search intelligence engine that gives our team the data they need to understand, analyze, and grow AI visibility, share of voice, and revenue. It tracks and measures the full impact of AI search, from prompt all the way to revenue, helping us drive the results that matter for our clients. Video Player PAVE not only expands on the foundational AI visibility measurement metrics, but it also includes new features developed from our team's needs and our clients' requests. The result is a tool built for our team with a focus on driving success for our clients. ### Here's how PAVE excels #### 1. Transparency around metrics Transparency around data calculations was a necessity for our team. Without understanding how a metric is calculated, we can't determine how meaningful it is to our clients' performance. With PAVE, we know exactly how every metric is derived, and we can explain and own it. When there's a discrepancy, we can trace it back to the underlying prompt, diagnose the issue, and move forward with confidence. #### 2. Control and flexibility Building our own AI search intelligence engine meant we could add features specifically designed to serve our clients better. One example is our ROI pipeline tracking, which paints a fuller picture of how AI search drives [website traffic](https://www.perrill.com/why-did-our-web-traffic-go-down/), how that traffic behaves once it reaches your website, and how technical [seo](https://www.perrill.com/seo-isnt-dead-9-tactics/) factors like structured data help optimize visibility in search results. In that same vein, we're able to continuously iterate on what we built, shaping PAVE around the data that we need to drive the best results for our clients and explain and own every metric as a trust signal for clients and for evaluating visibility in search results. And if any errors arise or if the software needs to be updated, our development team can handle that quickly in house through tailored development solutions to ensure minimal disruption to reporting, and if discrepancies appear, we can trace them back to the underlying prompt and review the context behind the output. Clear, direct answers and credible cited sources are what answer engines tend to reward, which is why we validate metrics so closely. Our team's responsiveness extends directly to PAVE. #### 3. Unified business intelligence Our team imagined a tool that could not only monitor common GEO metrics like AI visibility and share of voice, but could also track AI search leads throughout the entire funnel to connect AI visibility directly to ROI while showing how different queries affect website traffic quality and why the on-site experience still needs optimization after acquisition. Here's how PAVE expands on AI visibility reporting: * Session-level analytics * Branded vs non-branded prompt segmentation * Sub-brand consolidation and roll-up reporting * Citation source intelligence * Google Search Console integration * CRM and CallRail integration * ROI pipeline measurement Technical SEO best practices still matter for generative AI search, because a clear technical structure improves discovery and indexing, and schema markup helps ai systems and ai models parse context, including in ai overviews. It's an all-in-one platform that unifies your marketing rather than fragmenting AI visibility into its own silo, giving you a clearer view of your brand's presence and the [links](https://www.perrill.com/should-i-link-to-other-websites/) influencing visibility. And instead of leaving you guessing on the results, PAVE helps uncover tangible ROI of GEO services for your brand, while our team continues building development solutions that help diagnose and fix technical issues quickly, including verifying your site in Search Console to surface errors faster. #### 4. Strategy and execution, combined Measurement only matters when you have a team who can interpret it and act on it. With PAVE, our team is intimately aware of how the data is collected and calculated. Unified reporting also helps monitor your brand's presence across multiple AI platforms, including AI Overviews and other [AI-powered search](https://www.perrill.com/a-marketers-guide-to-ai-powered-search/) environments where AEO aims to appear in search results and voice assistants. We not only interpret the data, we report, strategize, and execute to create helpful content around user intent, user questions, and the needs of your audience. Because AI models and AI systems respond differently to prompts and queries, unified measurement matters when teams are building practical solutions, not just dashboards. * Track visibility across AI surfaces and prompt variations * Compare brand mentions, citation frequency, and links as practical indicators of share of voice and authority * Turn findings into content and optimization priorities Monitoring LLM answers is difficult because responses can shift based on memory and prior interactions, which makes centralized tracking especially valuable. We're not here to sell PAVE as a SaaS tool. We're here to use PAVE to power our GEO clients, getting them the results that matter. ### How clients will experience PAVE PAVE was born from a curiosity about how AI search actually works and a desire to drive better results for our clients. So determining how clients should interact with these insights was an important piece of the puzzle. PAVE captures a depth of data that most reporting platforms simply can't handle. Rather than simplifying it to fit an existing reporting tool, we built client dashboards accessible through individual logins. The result is a dedicated dashboard for clients to access their data, plus practical resources and tools that help them discover what matters without losing any of the depth that makes PAVE valuable. The dashboards highlight performance and progress across LLMs, and our team is always there to walk through what the data means and what we're doing about it. We not only interpret the data, we create solutions based on users intent, user questions, and audience needs, and share tips on what to publish. Answer engine optimization depends on direct answers to user queries and intent-driven, highly structured content. We use that insight to create helpful, conversational content instead of relying on static, keyword-based material. You'll know firsthand the insights we're rooting our recommendations in. ### AI visibility and ai driven search, powered by PAVE While we're officially rolling out PAVE across our GEO services, we're not done yet. And we probably never will be. PAVE is a continuous project, one that we'll continue to iterate on and grow to best fit our clients' evolving needs, with dashboards accessible through individual logins plus tools and resources to understand performance across LLMs and how answer engine optimization [structures content for LLMs](https://www.perrill.com/essential-steps-to-make-your-website-llm-ready/) like ChatGPT. That reporting also helps teams create highly digestible content and structure information so AI can easily discover and extract core facts. Ready to start driving better results for your business with [generative engine optimization](https://www.perrill.com/generative-engine-optimization-services/)—results you can track all the way through the pipeline, with practical recommendations based on how users, ai agents, search agents, customers, and ads surface information? [Reach out to get started](https://www.perrill.com/contact/).