AI reputation is not a number. It's a composite — five dimensions that tend to move independently and that benefit from being measured independently.
A brand can have flawless accuracy and hostile sentiment. Or favorable framing and missing completeness. Or strong consistency across four engines and one outlier that fractures the picture. A single score often hides where the damage actually lives.
The five dimensions are how 5W's Reputation Index audits AI-held reputation across ChatGPT, Claude, Gemini, Perplexity, and Google AI Overviews. Each dimension answers a different question. Each typically demands a different fix.
Accuracy — does the model state correct facts?
The most basic question. Is the founding year right. Is the leadership named correctly. Are the products described as they exist, not as they once did. Are the financial figures current.
Accuracy tends to fail for predictable reasons: out-of-date training data, low-authority sources getting weighted too heavily, scraped errors propagating across the citation graph. The fix typically runs upstream — push authoritative, current statements into sources the engines weight.
Sentiment — is the framing favorable, neutral, or hostile?
Sentiment is not about whether the answer "feels" positive. It's about which narrative the model summarizes when given a category prompt. "Best CRM platforms" — does the engine include the brand in the consideration set, or in the cautionary footnote?
Sentiment often skews when one critical article or one viral incident becomes the model's dominant source. Repair typically requires displacing that source's weight — not deleting it.
Completeness — are the key strengths cited?
Models compress. They tend to pick three to five attributes. If those attributes are the weakest ones — slow growth, recent layoff, a single product line in a multi-product company — the buyer often never sees the rest.
Completeness gaps are usually the result of a thin authority stack. The strengths exist but were never written into the sources the model pulls from at sufficient density.
Consistency — does the answer hold across engines?
Five engines, often five answers. ChatGPT and Claude train on overlapping but distinct corpora. Perplexity retrieves live and tends to bias toward recent sources. Gemini weights Google's index heavily. Google AI Overviews synthesizes inside the search experience.
A brand can read well on three engines and badly on two. Inconsistency is its own reputation risk — buyers cross-check, and a fractured picture often reads as untrustworthy.
Control — can the brand influence the answer?
The dimension everyone wants to be highest. Control is the brand's ability to shape future answers through earned media, Wikipedia edits, structured owned content, and the citation graph the model retrieves from.
Control is never absolute. Engines retrieve from the open web — not from a brand's preferred source list. But control is measurable: how much of the model's cited source pool is influenced by the brand's communications work, and how much is inherited from sources the brand cannot edit.
How the five combine
A composite score across the five dimensions gives a directional view. The dimension breakdown gives the repair plan. [Read: How to Audit What Every Major AI Engine Says About Your Brand]
A brand with high accuracy, high consistency, and low control is often one viral incident away from a reputation collapse it cannot easily fix. A brand with low completeness and high sentiment is typically leaving share on the table.
The composite tells you whether you have a problem. The dimensions tell you where to spend.
See also: Signals That Move AI Reputation · AI Reputation Glossary
No communications firm can guarantee specific outputs inside third-party AI systems. The discipline is shaping the inputs the engines retrieve from — not directing the engines themselves.
