What Magnent's Own AEO Optimization Agency Audit Revealed
We ran our own AEO audit process on ourselves before ever taking it to a client, and it surfaced gaps we didn't expect. Here's exactly what we found and what four weeks of fixing it actually changed.
Magnent's entity card appeared reliably in Gemini queries but inconsistently in ChatGPT, and was absent from Perplexity responses entirely.
A scenario that plays out often in the marketing industry: an agency builds a rigorous diagnostic process for clients, then applies a completely different, usually thinner, set of standards to itself. Magnent, an AEO optimization agency based in India, decided early on that this pattern was not worth repeating. Before taking the audit methodology to client accounts, the team ran the full process on Magnent itself.
In short, Magnent's internal audit surfaced three active citation gaps, two schema fixes that produced measurable improvement within four weeks, and a content freshness issue suppressing entity recognition across multiple AI engines. The results shaped how Magnent now structures client onboarding — and demonstrated that even an agency built around AI visibility accumulates blind spots when moving quickly.
What the AEO Audit Examined
The AI visibility audit process Magnent applies to clients covers four layers: entity recognition, citation sources, content structure, and query coverage. Applied internally, the diagnostic takes roughly thirty minutes to complete, but the remediation work the findings generate can run for several weeks.
The internal audit ran across three AI engines (ChatGPT, Gemini, and Perplexity) using seventeen query prompts covering variations of the agency's primary service terms, including "AEO optimization agency," "AI visibility agency India," and comparison queries such as "best AEO agencies in India for B2B brands." Each prompt was tested across multiple sessions to account for model variability.
What the AEO Optimization Agency Audit Found
Entity Recognition Was Inconsistent Across Engines
Magnent's entity card appeared reliably in Gemini queries but inconsistently in ChatGPT, and was absent from Perplexity responses entirely. The root cause was a mismatch between how Magnent described its services on its own website and how third-party sources referenced the agency. Schema markup was present but incomplete: the Organisation schema was missing service categories that aligned with the query language AI engines associate with the AEO space.
This is a pattern Magnent sees consistently across client audits. Brands operate with a strong internal vocabulary for what they do, but AI engines pull from third-party descriptions as much as from the brand's own site, and when those descriptions diverge, entity recognition becomes unreliable.
Citation Sources Were Concentrated in Two Domains
Of the queries where Magnent appeared in AI-generated answers, the majority of citations came from a narrow cluster of third-party sources. This concentration creates structural fragility. Research on how AI engines prioritise content in their retrieval systems finds that a small number of high-authority domains carry disproportionate citation weight, a dynamic that McKinsey's State of AI research{:target="_blank" rel="noopener"} has examined in the context of how enterprise information systems weight source authority (McKinsey Global Institute, 2025). The AEO and AI visibility trends data tracking Indian brands documents how citation source concentration plays out in practice, with Perplexity showing the sharpest volatility when any source within that cluster shifts its coverage.
The fix involved expanding third-party coverage: securing accurate representation on additional comparison platforms and external publications, and ensuring consistent service descriptions across all external surfaces.
Content Freshness Signals Were Stale on Key Pages
Two high-value pages on the Magnent site, both covering core service offerings, had not been updated in several months. AI engines treat content freshness as a relevance signal, and for an agency whose value proposition is tied to a fast-moving category, stale pages carry a double penalty. They signal both outdated information and a gap between the agency's claimed expertise and its visible activity.
Both pages were updated with current observations, recent context from client work described in general terms, and revised internal linking structures.
What Changed After the Fixes
Four weeks after schema corrections and source diversification work was completed:
- Perplexity citation rate moved from zero to measurable across five tracked queries
- ChatGPT entity recognition became consistent across the seventeen-query test set
- Gemini citations held at previous levels and improved on comparison-style queries
These results align with what Magnent observes across client engagements: schema and source fixes produce the fastest signal changes, while content freshness improvements take slightly longer but tend to produce more durable gains.
What This AEO Optimization Agency Audit Confirms for Indian Brands
| Issue Category | Frequency in Internal Audit | Typical Timeframe to Fix |
|---|---|---|
| Schema incompleteness | High | 1–2 weeks |
| Citation source concentration | High | 3–6 weeks |
| Content freshness gaps | Medium | 2–4 weeks |
| Query coverage gaps | Medium | Ongoing |
The non-obvious finding from running this process internally: agencies and professional services firms face a structurally different AEO challenge from product companies. The queries that matter are dominated by comparison intent, such as "best AEO optimization agency in India," rather than informational intent, and comparison queries depend almost entirely on third-party coverage quality rather than on what the agency says about itself.
Brand-side content, however well-structured, does not substitute for trusted third-party references when AI engines construct a comparative answer. For Indian B2B brands reviewing whether to invest in AEO work, the Magnent FAQ addresses common questions about how the engagement process works.
FAQ
What does an AEO optimization agency audit check? An audit checks how and whether AI engines recognise, cite, and describe a brand when users ask relevant questions. The process covers entity signals, schema markup, citation source quality, content structure, and query coverage — the components that determine whether a brand appears in AI-generated answers at all.
How long does it take to see results from AEO optimization work? Schema and source fixes typically produce measurable changes within four to eight weeks. Content freshness improvements take slightly longer. Query coverage expansion is an ongoing process rather than a fixed-term project.
Can an AEO agency audit itself objectively? The methodology applies to any entity, including the agency running it. The value of Magnent having done this is not objectivity — it is calibration. The team now knows exactly what the audit surface looks like from the inside, which improves how findings are communicated to clients and how remediation priorities are set.
What is the most common finding in AEO audits of Indian brands? Based on client engagements, the most consistent finding is citation source concentration combined with schema incompleteness. Both are fixable, and both tend to have the highest citation impact when addressed together.
Why does Perplexity behave differently from ChatGPT in AEO audits? Perplexity functions more like a retrieval-based engine than a trained-knowledge engine. Its citations depend heavily on live third-party sources, making it more volatile than ChatGPT when those sources shift coverage. This is why Perplexity-specific source strategy is treated as a distinct component in an AEO optimization plan.