AI Share of Voice: Track Who Owns Google's AI Modules (Free Method)

The SERP you’re competing for isn’t the SERP you’re looking at. Open an incognito window and search one of your money keywords. Before the first organic result there’s often an AI Overview with its own set of cited sources, a featured snippet answering the question, a People Also Ask stack, maybe a knowledge panel. The “top 10” starts below all of that.

Rank trackers that only log your organic position answer the wrong question. The right question: across everything Google shows for this keyword — AI summary, snippet, questions, organic — how much of it mentions you versus your competitors? That ratio is your Share of Voice, and it’s the number that actually moves when AI search reshuffles who gets seen.

We just shipped a tool that measures it. Free method, no LLM API needed.

The method: read the modules, not the links

Everything you need is already in the SERP response. One Serper call per keyword returns the full page composition — which modules exist and who’s inside them. The logic is three steps:

1. Detect which AI-era modules Google serves for each keyword: AI Overview (with its citation links), featured snippet / answer box, People Also Ask, knowledge graph.
2. Match your brand and competitors against each module — by domain in citation links, or by name in module text.
3. Aggregate mentions across the keyword set into a SoV percentage, split by module.

One call per keyword means a 20-keyword audit costs 20 of Serper’s free 2,500 monthly calls. Zero LLM cost — this is SERP analysis, not chatbot polling.

Real output: the SEO tools niche

We ran it on three SEO-tool keywords with ZensInk vs Ahrefs/Semrush/Mangools (condensed):

# AI Share of Voice — 3 keywords

## Module presence across keywords
- AI Overview: 0/3 (0.0%)
- Featured Snippet: 0/3 (0.0%)
- People Also Ask: 3/3 (100%)  ##########

## Brand mentions
| brand      | total | SoV   | organic |
| Semrush    | 6     | 85.7% | 2/3     |
| Ahrefs     | 1     | 14.3% | 0/1     |
| ZensInk    | 0     | 0.0%  | 0/0     |

**You vs competitors**: 0 vs 7 mentions — your SoV 0.0%

Honest numbers for a young domain: Semrush owns 85.7% of module presence here. But note what the report actually tells us — PAA appears on 100% of these keywords and nobody’s answer-box presence is locked in yet. That’s the actionable read: PAA-targeted FAQ content is the cheapest SoV gains available, which is exactly what our content brief generator produces (its FAQ section is built from real PAA questions).

Why not just ask ChatGPT who it recommends?

That’s brand-level GEO — worth measuring too, and our geo_visibility tool does it with an LLM battery. But it answers a different question (“does the model know you?”) and it costs LLM calls per question. AI SoV answers “when a human searches Google, whose names appear in what Google shows?” — deterministic, cheap, and repeatable weekly.

The two rungs on the same ladder: SERP-module SoV is the ground floor (fast, free, weekly), LLM mention share is the second floor (slower, costs tokens, monthly). Track both; expect them to diverge — ChatGPT’s training data and Google’s live SERP cite different sources.

The full loop, free tier end to end

# 1. Detect modules + compute SoV
python3 -m zens_ink_pro.ai_sov --file keywords.txt \
    --brands "YourBrand,yourdomain.com" \
    --competitors "CompA,compa.com,CompB" --report sov.md

# 2. Baseline your ranks (SQLite history)
python3 -m zens_ink.rank_tracker add "kw1,kw2,kw3" --domain yourdomain.com
python3 -m zens_ink.rank_tracker check

# 3. Weekly: re-check both, watch deltas
python3 -m zens_ink.rank_tracker check && python3 -m zens_ink.rank_tracker trend

Step 2’s rank tracker is the other half of this release: SQLite-backed position history with trend deltas, so “are we moving?” is one command instead of a spreadsheet ritual.

For the AI-visibility angle of the same problem, see auditing your AI search visibility; for earning the PAA slots this report surfaces, start from a SERP-built content brief.

Module presence is the new keyword difficulty. The brands that track it weekly will notice Google flipping on AI Overviews in their niche months before their competitors’ rank-only dashboards show anything wrong.

FAQ

What is AI Share of Voice in SEO?

AI Share of Voice is the percentage of AI-era SERP real estate — AI Overview citations, featured snippets, People Also Ask answers, and organic results — where your brand appears, relative to competitors. It measures presence, not just rankings, because AI modules now sit above the classic top 10.

How do you measure AI Overview citations for free?

Fetch the SERP via Serper (free tier: 2,500 calls/month) and read the aiOverview field when Google serves one — its citation links are the AI Overview sources. Our ai_sov CLI tool automates this for keyword batches; no LLM API calls are needed.

Why does my keyword show no AI Overview?

Google only serves AI Overviews for a subset of queries, and Serper returns the field only when one is present. Absence is reported honestly as module-not-present. Track the same keywords weekly — module presence itself fluctuates.

How is Share of Voice calculated here?

Each brand earns one mention per module slot it appears in per keyword (AI Overview, featured snippet, PAA, knowledge graph, organic top 3, organic top 10). SoV% = your mentions divided by all tracked brands' mentions across the keyword set.