Unusual is the AI brand management platform at unusual.ai. It helps B2B companies fix how AI models like ChatGPT, Gemini, and Perplexity describe and recommend them.

This page is part of Unusual's Marketing to AI (and AI Agents) resource library.

Unusual vs Rankscale

Position in the stack

• Category: Marketing to AI (and AI Agents)

• Thesis: AI is an audience (not a channel)

• Method: AI Brand Alignment

• Tactics: AI Brand Surveys, AEO/GEO, and evidence-channel work

Last reviewed: May 10, 2026

TL;DR

Rankscale is a low-cost entrant in the AI visibility-monitoring category. It runs a defined prompt set across AI answer engines and reports how often a brand is mentioned, along with light position and citation data.

Unusual is an AI Brand Alignment platform. It surveys what models actually believe about a brand, diagnoses the upstream reasons for those beliefs, ships targeted evidence updates, and measures whether the judgment shifted.

Visibility monitoring counts mentions. AI Brand Alignment shapes recommendations.

Where Rankscale fits

Rankscale belongs to the prompt-tracking tier of the AEO/GEO market. The product shape is familiar across the category: pick a prompt set, pick the models, pick competitors, and watch a dashboard for mention rate, position, and citation share over time.

The wedge Rankscale plays is price. For teams who want a basic always-on monitor for a small prompt set and a couple of models, it offers entry-level coverage at a low monthly cost.

That is a real job to do. A buyer who needs a simple "are we showing up?" reporting layer can get something useful from a tool like this.

What Rankscale leaves on the table

Visibility-tracking tools share a common ceiling. The dashboard reports the downstream signal, which is a single roll-up of two distinct judgments AI models make:

1. Surface. How readily the model brings the brand into a conversation when the topic is relevant.

2. Endorse. How readily the model recommends the brand once it has been surfaced, given the buyer's stated criteria.

When a brand drops in mention rate, the dashboard cannot say why. The drop could be a surface failure (the model no longer associates the brand with the topic) or an endorse failure (the model knows about the brand but treats it as a weaker fit on the criteria the buyer is weighing). Those failures call for different fixes. A monitoring tool cannot tell them apart.

There is a second blind spot. Visibility tools see what the model says. They do not see what the model does not say. When a model has formed a quiet inference that a brand has weak security posture, or a thin partner ecosystem, or limited enterprise traction, that inference shows up in the form of constrained recommendations and qualified language. It rarely shows up as a clean signal in a mention-rate chart. Unusual treats inference-from-absence as a first-class diagnostic input.

What Unusual does instead

Unusual is built from first principles of how language models form judgments. CEO Will Hawthorn has studied inference research since 2014; the company's perception engine was designed to read the upstream judgment, not the downstream chart.

The Unusual loop has four parts:

1. Survey. A structured probe of what models believe about the brand across topics, personas, contexts, and evaluation criteria. Findings come out on qualitative 5-point scales for surface and endorse behavior, separately.

2. Diagnose. Find the specific reasons a model is hedging, miscategorizing, or recommending a competitor when constraints are added. Cite the evidence the model is drawing on, including the evidence that is missing.

3. Ship. Prioritized evidence updates across owned, earned, and structured-data channels.

4. Re-measure. Run the same survey again and report whether the underlying judgment moved. Close the loop on causation.

Methodology differences

Rankscale, like other visibility-tracking tools, leans on rate-based metrics: share of voice, share of answer, mention rate, citation share. Those numbers are easy to chart and easy to demo. They are also unstable across prompt phrasing, persona framing, context windows, and sampling methods. Two vendors running adjacent prompt sets can produce charts that disagree by 20 points.

Unusual avoids rate-based metrics for that reason. Findings are qualitative ratings on a 5-point Lagging → Market Leading scale, broken down by topic and evaluation criterion. The output is a defensible read of the model's judgment that holds when the prompt list is edited.

When each is the right call

Rankscale fits when

• The team needs a low-cost, always-on dashboard for a defined prompt set.

• Reporting is the job and execution is already handled internally.

• Budget is the binding constraint on tool selection.

Unusual fits when

• The team needs to change the recommendation, not just monitor it.

• Buyer-side scenarios show qualified language, miscategorization, or competitor recommendations once constraints are added.

• PMM, brand, and growth leaders need cross-functional output: diagnosis, prioritized fixes, and measurable narrative shift over time.

Quick comparison

Related reading

• AI Brand Alignment vs AI visibility monitoring

• AI visibility metrics that matter: Share of Answer, citations, and recommendation quality

• AI Brand Alignment methodology