How Aelo works: measurement, diagnosis and a step-by-step plan

Below, we describe each of these four steps: where the data comes from, what it reveals, and what you end up with.

If you’re new to this topic, read the guide first Brand visibility in AI: how it works and how to measure it.

Measurement: how do we check whether models are switching brands?

You start with phrases that describe what your customers are looking for, such as „fertiliser lime manufacturer” or „accountancy firm for companies in Katowice”. The system translates these into the kinds of questions people ask in chatbots and asks them without mentioning your brand name. This is what it looks like neutral scan: we’re checking whether the model can name the brand itself, rather than whether it recognises it when asked. We measure phrases containing the brand name separately, if you include them.

The questions are sent to four models: ChatGPT, Claude, Gemini and Perplexity. The measurements are taken daily, at the same frozen query set, because AI responses change almost every time the system is run, and a single result proves nothing. The text explains why this is the case Why does AI give a different answer every time?.

From the assessment that marks the start of your work at Aelo, you’ll receive:

• in how many responses does the brand appear, and in which models,
• Does the model simply replace it (reference), or whether it cites your website as the source (citation),
• alongside which competitors the brand appears and who takes its place,
• how models talk about her: positively, negatively or neutrally (sentiment),
• a full report once a month, which compiles the results from all the scans.

Details of how the results are calculated are set out in our methodology, and the page module itself Visibility.

Diagnosis: how do we know what is preventing the recommendation?

The measurement results are sent to Synapsis, that is to say knowledge graph About the brand. We build this automatically from several sources: indexing of your website, scans, citations in model responses, competitor analyses, audits and implemented changes. Another source is the Brand Identity form, which you complete after registering in the brand settings.

This allows you to spot discrepancies that the measurement alone won’t reveal. For example, the fact that the model recognises the name of your product but doesn’t link it to the company (brand aliases). Or that they recognise the brand but do not recommend it nonetheless, because there are no independent sources to corroborate it. We describe this threshold in the text Knows the brand, but doesn’t recommend it and in the entry assessment and recommendation.

Plan: what do you get after the assessment?

Based on the diagram, we draw up specific actions in Tickets. Each one relates to a single task: adding a fact to the page, correcting a service description, or updating data that the models interpret differently. Paid plans also include Outreach, which covers activities outside your website: where to set up a company profile, which directories and websites the brand should appear on, and who is worth reaching out to.

The following are used to work with the plan Advisors, namely the four advisers: Consultant, Analyst, Planner and Strategist. They use the same data as the rest of the platform, so their recommendations are based on facts about your brand, rather than general advice.

Activities: what sort of materials does the Studio produce?

Studio prepares the materials needed to carry out the plan: website content and graphics, and, for higher-tier plans, advertisements as well. The materials are created in line with the brand guidelines, so they stick to the facts about the brand. We charge for the use of Studio in AeloCoins.

Aelo doesn’t publish anything on its own. We prepare the content, and you check it, correct it and either upload it to the website or send it to the client. What ends up being published is entirely up to you.

How does this work in practice?

The „Węże” lime works in Działoszyn, a lime producer, had an extensive website with several dozen subpages. In June 2026, none of the models tested mentioned the brand in their answers to 18 questions about its category.

• Measurement Showed 0 out of 18 queries containing the mention and 6% page citations.
• Diagnosis She pointed out that the models recognise the name of the flagship product, but treat it as a type of lime rather than a specific company’s product, and that the phrase linking the product to its price and manufacturer yields the weakest result.
• Plan It included changes to existing content: a clear statement of compliance with the standard, a table of specifications, a section on price, and the „manufacturer” label on the weakest local subpages.
• Function was carried out by the client. Not a single new subpage was created.

In September, the brand appeared in 18 out of 18 search queries; the models cited its website in 94% phrases, and the first model began to recommend it. You can find a full description in the two parts of the case study: From 0 to 17 mentions on AI in five weeks and An AI model recommends a brand for the first time. He explains why standard SEO wasn’t enough here Visibility in AI and SEO.

Who is Aelo for?

Aelo is designed for marketing and SEO agencies looking to manage their clients’ AI visibility, as well as for companies wishing to manage it themselves. Agencies can manage multiple brands within a single account and, on the Custom plan, utilise the API and white-label version. Further details can be found on the website Aelo for agencies and Aelo for small businesses.

How do I get started?

The easiest way to try out Aelo is with the Free account: one brand, measurements across four models and a full report – no payment card required. We explain exactly what this includes in the article Free brand visibility measurement in AI: Free account.

Try it for free

FAQ

Sources

• Aelo reports for Wapiennik „Węże” dated 24 June, 27 July and 10 September 2026 (published with the client’s consent)