A Five-Step, Documented Method for
AI Product Discovery
ACTRISE follows one evidence-led process for every engagement: Measure, Diagnose, Prioritize, Improve, Re-measure. Each step produces a written record so you can see exactly what was tested, what was found and what changed.
This is not a black-box audit. It is a reproducible method built for growth-stage Indian D2C brands who want to understand their position in AI-powered buying journeys before deciding what to do about it.
Five Steps, Five Sets of Deliverables
Every ACTRISE engagement runs through the same sequence. Each step feeds the next, and every step closes with documented outputs you can read, question and keep.
Measure: Controlled Queries Across ChatGPT, Gemini and Perplexity
We start by defining up to 3 priority products and a set of 12 controlled buyer queries that reflect how real customers might ask about your category. These queries are run across ChatGPT, Gemini and Perplexity under documented conditions, producing up to 36 baseline observations.
We also run the same queries against up to 3 named competitors, so you see not just whether your products appear, but what appears instead when they do not.
- Query Map: the exact buyer questions used and why they were chosen
- Baseline Observations: what each AI platform actually returned, recorded as-is
- Competitive Snapshot: how up to 3 competitors show up on the same queries
Diagnose: Comparing Observations Against Product Facts and Competitors
Raw observations alone do not tell you what to do. In the diagnosis step, we compare what the AI platforms said against your actual product facts, your site content and how competitors are being represented on the same queries.
We mark where the gap is clear and where it is uncertain. AI platform outputs can vary between runs and over time, and we document that variability rather than smoothing it over.
- Illustrative Diagnostic Findings: where your products are underrepresented, misrepresented or absent
- Clear separation between supported findings and open questions
- No inflated claims about cause and effect between any single factor and AI output
Prioritize and Improve: Ranked, Agreed, Implemented Actions With a Change Log
From the diagnosis, we build a Prioritized Action Plan: a ranked list of possible improvements, scored by relevance to the observed gaps and feasibility given your current site, content and product data. You review and approve the plan before anything changes.
Once agreed, we implement up to 5 bounded actions, subject to the access and approvals you provide. Every change is recorded in a Change Log so there is a clear before-and-after record tied to the original diagnosis.
- Prioritized Action Plan: ranked by relevance and feasibility, not guaranteed results
- Implementation Pack: the agreed changes, described in detail
- Change Log: a dated record of exactly what was modified and when
Re-measure: Repeating the Checks and Reporting What Changed
After implementation, we run one controlled re-measurement cycle using the same query set and platform conditions established in step one. This keeps the comparison fair and lets you see what moved and what did not.
The Re-measurement Report sets the new observations side by side with the baseline and states plainly what changed, what stayed the same and what remains uncertain. We do not promise placement, rankings, traffic or sales. We report what the evidence shows.
- Re-measurement Report: new observations compared directly against baseline
- Executive Summary: a plain-language account of what was learned
- Honest treatment of uncertainty: no overstated attribution of results
Questions About the Method
What brands typically ask before starting a consultation.
What does the Measure step involve?
What does the Diagnose step produce?
How many actions are implemented in the Improve step?
What does the Re-measure step report?
Is this a guarantee of improved AI representation?
How long does the full engagement take?
Request a Consultation to Walk Through the Method
If you want to see how this method would apply to your products, categories and competitors, the next step is a short discovery call. We will walk through the five steps against your specific situation and answer any questions about scope, timeline and deliverables.