The context
The Indian TWS market is crowded and price-led, and buyers rarely arrive with a brand in mind. They ask for the best earbuds under ₹2,000, the best pair for running, the ones with ANC that actually works. Nine brands compete for a shortlist that holds three to five names.
This brand had distribution and recall. In AI answers it was a footnote.
The challenge
Fourth of nine by share of voice — and near-last by recommendation rate.
The Beacon audit covered the category’s high-intent queries across ChatGPT, Gemini, Claude and Perplexity, split across price, feature, use-case and brand-trust intent.
The brand ranked fourth of the nine players tracked on share of voice, and its recommendation rate sat well behind that — a brand the engines named in passing rather than put forward. Its own domain was almost absent from the citations behind category answers, and in a handful of queries it did not surface at all.
The pattern underneath the numbers mattered more than the numbers. Price queries carried the brand. Feature queries, use-case queries and brand-trust queries belonged to three larger competitors, and those are the queries where the buyer has already decided to spend and is choosing whom to spend it with.
The strategy
Pick the narratives, abandon the rest
The brand committed to a small set of narrative themes — the ones where its actual product strengths matched an unowned query. Everything else was left alone for the period of the engagement.
Build the evidence a verdict needs
Engines do not recommend on adjectives. Each narrative was backed with measurable claims — battery drain tests, ANC measurements, latency numbers — published as text tables inside third-party reviews so an engine could lift the figure and cite the source.
Publish where the category is already being answered from
Placement followed the citation data from the audit: the Tier-1 and Tier-2 Indian tech titles that already appeared as sources in category answers.
Hold the format standard on every piece
Keyword-matched headline, comparison table, price-band structure and an explicit verdict. A review that ends on ‘depends on your needs’ contributes nothing to an engine assembling a recommendation.
The execution
Over four months the programme commissioned buying guides and head-to-head comparisons across the publishers that already surfaced as citations in category answers. Those two formats do most of the work in ‘best X under ₹Y’ responses.
Content was structured to the price bands the category actually indexes on: ₹1,000, ₹1,500, ₹2,000, ₹2,500. Where a piece carried a verdict for a band, the brand was positioned with a specific reason rather than a general endorsement.
Beacon re-ran the same query set monthly, tracking movement per engine, and coverage was reallocated toward the engines and themes where the shift showed first.
The outcome
The recommendation rate was the number that moved — from a low double-digit baseline to 60% over four months, on the same query set and engines. Share of voice rose more modestly to 25%, which is the expected shape when a programme is aimed at conviction rather than reach. The brand ranked first on both measures against its tracked competitors in those categories.
Baseline and closing figures both drawn from Beacon audits of the same category query set across four AI engines.
What we learned
A mid-tier brand cannot outspend the category leaders across every query, and it does not need to. The shortlist an engine returns is short. Owning four narratives completely is worth more than a thin presence across 10.
Price-anchored visibility is the most fragile kind. A brand that only surfaces when a budget is named in the prompt loses the moment the buyer asks about a feature instead.
- AI Visibility Audit (Beacon)
- Narrative prioritisation
- Publisher network strategy
- Comparison and buying-guide commissioning