# Heatseeker Heatseeker is an AI-powered customer intelligence platform. > Heatseeker helps brands build customer understanding they can put to work across the business. Customer-specific AI personas (synthetic personas) make that understanding available when teams need it, and live market experiments supply behavioral evidence that helps calibrate and strengthen them. The Customer Understanding OS is the reusable customer-intelligence layer. We work with enterprise marketing and insights teams globally; our teams are based in the US and Australia. ## Pages - [Platform](https://www.heatseeker.ai/platform): Customer-specific AI personas, live market experiments, how customer understanding is applied to decisions, and the Customer Context Layer. - [Customers](https://www.heatseeker.ai/customers): Named results from brands including DoorDash, L'Oréal, KPMG, Hearst, Cisco and Tyro. - [Pricing](https://www.heatseeker.ai/pricing): Project and annual platform plans (Project, Base, Pro, Enterprise), priced to the decision. - [Security](https://www.heatseeker.ai/security): SOC 2 Type II, AES-256 encryption, data residency across the US, Australia and the EU, and lookalike brand testing. - [Insights](https://www.heatseeker.ai/blog): Articles and the Finding the Heat podcast on behavioral ground truth. - [Contact](https://www.heatseeker.ai/contact): Direct inboxes, both contracting entities, and our offices. - [Careers](https://www.heatseeker.ai/careers): How the team works, a week at Heatseeker, and open roles. - [Book a demo](https://www.heatseeker.ai/demo): A 30-minute working session on one real decision. ## Support - [Support center](https://www.heatseeker.ai/support): How a market test works, how an experiment is designed and run, how to read your results, and setup guides for connecting ad accounts and MCP clients. - [How a market test works](https://www.heatseeker.ai/support/how-a-market-test-works): Real people in your segment see real ads, split at random into equal cells and scored on what they do; why non-responders are part of the measurement, why a week is enough, how it compares to surveys, panels and A/B testing, and how results are read against baselines and benchmarks. - [How to read your results](https://www.heatseeker.ai/support/how-to-read-your-results): The winner card, the performance index, the four conviction ratings, expected loss and decision confidence, and the by-segment heatmap. - [Connect to Heatseeker with MCP](https://www.heatseeker.ai/support/connect-to-heatseeker-with-mcp): Connect Claude, ChatGPT, Copilot or any MCP client to a Heatseeker workspace; server URL, OAuth, the 13 available tools, troubleshooting. - [Connect Meta to Heatseeker](https://www.heatseeker.ai/support/connect-meta-to-heatseeker): Share Meta Business Suite pages and ad accounts with Heatseeker, or connect a self-serve account. - [Connect LinkedIn to Heatseeker](https://www.heatseeker.ai/support/connect-linkedin-to-heatseeker): Add Heatseeker as a LinkedIn Business Manager partner and share pages and ad accounts. - [Connect TikTok to Heatseeker](https://www.heatseeker.ai/support/connect-tiktok-to-heatseeker): Share your TikTok account and ad account with Heatseeker by QR code or from TikTok Business Center, or connect a self-serve account. - [Enable Meta ad performance data](https://www.heatseeker.ai/support/enable-meta-ad-performance-data-for-heatseeker): View-only Meta ad account sharing so Heatseeker AI can read performance. - [Enable LinkedIn ad performance data](https://www.heatseeker.ai/support/enable-linkedin-ad-performance-data-for-heatseeker): Viewer-role LinkedIn ad account sharing so Heatseeker AI can read performance. ## Guides - [How do you know your AI is right? For sure?](https://www.heatseeker.ai/guides/ai-market-research): What separates real AI market research from a confident guess, a reality check against live experiments, forward-looking signal, and a private model. The category explainer. - [Synthetic personas will tell you anything. Here's how to know when they're right.](https://www.heatseeker.ai/guides/synthetic-personas): What a synthetic persona is, the tools that use them, and the two questions that decide trust, what it's grounded on and what it's validated against. - [What people say. What people do. The gap that eats launches.](https://www.heatseeker.ai/guides/say-do-gap): Why stated intent predicts real buying only 20 to 30% of the time, which research tools inherit the gap, and how behavioral methods close it. - [Sometimes the fastest way to sell more is to charge more](https://www.heatseeker.ai/guides/willingness-to-pay): Why willingness to pay defies surveys, and how to find the real number by watching what buyers do. - [B2B market research for teams that need proof, not panels.](https://www.heatseeker.ai/guides/b2b-market-research): How to validate B2B positioning and messaging against real buyer behavior instead of panel opinion. - [B2C market research that predicts what shoppers actually buy.](https://www.heatseeker.ai/guides/b2c-market-research): Behavioral methods for consumer brands, and why stated preference misses. - [B2C customer segmentation that holds up in-market.](https://www.heatseeker.ai/guides/b2c-customer-segmentation): Building segments from revealed preference so they survive contact with a real campaign. - [Business validation: prove the demand before you defend the budget.](https://www.heatseeker.ai/guides/business-validation): Testing demand for a new business, product or market before committing spend. - [Product and innovation leadership: order the roadmap by likelihood of success.](https://www.heatseeker.ai/guides/product-innovation): Using behavioral evidence to sequence a roadmap. - [SaaS conversion rate optimization starts with the message.](https://www.heatseeker.ai/guides/saas-conversion-rate-optimization): Why message-market fit outranks funnel tweaks, and how to test it. ## Research - [Understanding customer needs: what our comparison of AI models and market responses found](https://www.heatseeker.ai/research/can-ai-predict-customer-behavior): A selected, retrospective study of 11 live-market experiments. ChatGPT, Claude and Gemini's top choices matched the recorded top performer 4, 3 and 1 times out of 11 (8 of 33 picks, same 11 tests). Method, results table, limitations, FAQ and CSV data (CC BY 4.0). - [Research index](https://www.heatseeker.ai/research): Original Heatseeker studies that compare AI output with observed customer response in live market experiments. ## Press and author - [Press](https://www.heatseeker.ai/press): Articles by Kate O'Keeffe in Forbes Technology Council and Harvard Business Review, plus coverage of Heatseeker. - [Kate O'Keeffe](https://www.heatseeker.ai/about/kate-okeeffe): CEO and co-founder of Heatseeker. Previously a Partner at BCG Digital Ventures and Head of Innovation at Cisco; co-author of two Harvard Business Review articles; member of the Forbes Technology Council. ## Articles - [Concept testing: what it's actually for.](https://www.heatseeker.ai/blog/concept-testing) - [The best idea in the room usually dies in the room](https://www.heatseeker.ai/blog/best-idea-dies-in-the-room) - [Marketing budgets are down 30%. The winners won't have the best AI stack.](https://www.heatseeker.ai/blog/marketing-budgets-ai-stack) - [Why surveys quietly compound into 0.1% confidence.](https://www.heatseeker.ai/blog/customer-research-model) - [How real and synthetic data work together to outlearn your market.](https://www.heatseeker.ai/blog/real-and-synthetic-data) ## Podcast - [Finding the Heat, Ep. 1: the CMO of L'Oréal on bridging the art and science of marketing.](https://www.heatseeker.ai/podcast/finding-the-heat-ep-1) ## Legal - [Privacy Policy](https://www.heatseeker.ai/legal/privacy-policy) - [Terms of Service](https://www.heatseeker.ai/legal/terms-of-service) - [Data Processing Addendum](https://www.heatseeker.ai/legal/dpa)