Know where your next marketing pound should go, and have the evidence to defend it.

We report every result with its margin. Each signal carries a confidence rating based on the sample size, recency and consistency of the evidence behind it. Impact's Bayesian model goes further, reporting the estimate, its range, and how much of that range your data supports. A wider range calls for a more cautious move, and that is reflected in every recommendation and in scenario planning. Where the evidence is thin, we flag it.

We treat it as an emerging signal rather than an established KPI. naras shows whether ChatGPT, Claude and Google’s AI mention, rank and recommend you when buyers ask, since more buyers now start there. We don’t claim a fixed link to sales, and AI answers change with each model update, so it works best as an early-warning indicator. It sits alongside funnel measures such as awareness and consideration, which have decades of evidence behind them.

Surveys are commissioned with real respondents through Cint, a major respondent marketplace, with questions customised to your brand and category. If you already run a tracker we can take it over and onboard your history, keeping your audience definitions, segmentation and markets so your trends stay consistent. It replaces the tracker rather than running beside it, so there’s no “paying twice” and no conflicting numbers. (The demo uses synthetic data, clearly marked.)

Synthetic Research is a tool for testing hypotheses. Its respondents are modelled on your workspace’s survey data rather than invented personas, and each question is run across three AI model families so that no single vendor’s biases determine the answer. Results are reported as ranges, and a published validation against held-out human survey data determines what each cut may claim; anything not yet checked is withheld, and we name it. Use it to narrow the options in minutes, then confirm the winner with real research. It does not replace fieldwork, and it does not produce brand-health figures.

Setup takes a short conversation, and we do most of the work. The search and AI audit is quick; a commissioned survey takes weeks (versus the months a tracker takes to stand up). For Impact, a spreadsheet of media and sales history is enough to begin.

Your data stays yours, encrypted in transit and at rest, including the sales data Impact ingests. Every brand gets its own isolated workspace, unreachable from another’s by construction: the application only resolves data for the workspace you’re in, and the database enforces the boundary again with row-level security, so even a coding mistake can’t return another client’s data. We’re glad to take your security or procurement team through the architecture before anything moves.

Yes. naras is built for it: each client brand gets its own workspace, kept separate, with its own evidence, signals and assistant. Exports are client-ready, so your team can drop them straight into a plan, a QBR or a pitch. Co-branded and white-label outputs are available on request. Talk to us about agency access.

Pricing depends on scope: which modules, how many markets, and how often you run. Tell us what you need and we will give you a clear price up front.

We’re early, and working with our first cohort of brands. We don’t borrow logos or invent results: the figures on this page are drawings of the product, and their numbers are illustrative. Ask us what we can show you.