What is the difference between the profile and a skill?

The profile decides voice and boundaries across everything. A skill decides structure and steps for one kind of output — a weekly summary, a competitor brief, a customer-facing note. They compose. Keep each skill narrow: one job, one shape.

Can I ask a skill for a share-of-voice ranking?

You can ask, but you will get an honest gap rather than a number. Competitor rates share your denominator and so do not sum to 1, and there is no brand rank metric in the product. Write skills around mention rate, citation rate, source share and average citation position instead.

If I downgrade, do I lose what I wrote?

On Starter the surface shows an upgrade state and the agent carries no stored brand context, so its answers become generic again. Nothing you have configured elsewhere in the workspace is affected by being on Starter.

Knowledge and skills

A brand profile the agent checks its own drafts against

Knowledge and skills gives the agent standing context about your brand and reusable named instructions for recurring jobs. It changes how the agent writes, not what it measures — mention rate, citation rate and source share are untouched by anything you put here.

Get Started Book A Demo

A brand profile
The context you would otherwise paste into every chat and every automation.
Named, reusable skills
One job, one shape: a weekly summary format, a competitor brief structure.
Rules, not adjectives
“Never call it a platform” does work. “We are innovative” does not.
It cannot invent a number
A skill asking for a metric we do not measure gets an honest gap.

By Karl-Gustav KallasmaaLast updated September 3, 2026Behaviour on this page is checked against docs.attensira.com

What it changes, and what it deliberately does not

Knowledge and skills changes how the agent writes. A brand profile tells it who you are, what you sell and how you talk. Skills give it reusable instructions to apply every time. Your mention rate, citation rate and source share are untouched by anything on this screen.

That separation is the point. Configuration that quietly moved your numbers would make the numbers worthless. Measurement is fixed and boring on purpose; voice is the part you get to configure.

Without it the agent still answers correctly. It just answers generically, and you end up pasting the same context into every chat and every automation — which is both a cost and a source of drift, because the fifth paste is never quite the same as the first.

The brand profile

The brand profile is the standing context the agent carries into every conversation in the workspace. Put in it the things you would otherwise repeat: what the product does, who buys it, the words you use for your own categories, the competitors you care about, and the claims you are not willing to make.

Be specific and be short. Vague positioning makes output vaguer. Concrete constraints do visible work on every answer that follows, because the agent can check a draft against them.

The best line in the documentation is the one worth stealing for your own first draft: write the profile as rules the agent can check its own draft against, not as marketing copy. “Avoid X” is actionable. “We are innovative” is not.

Category vocabulary — the words you use, and the ones you refuse.
Claims you will not make, which is usually the highest-value section.
Who buys, so examples are pitched at the right reader.
Competitors you care about, so a brief comes back about the right set.

Skills

A skill is a named, reusable instruction for a recurring job: a weekly summary format, the structure of a competitor brief, the tone for a customer-facing note. Instead of restating the format every time, you point the agent at the skill.

Skills compose with the brand profile. The profile decides voice and boundaries; the skill decides structure and steps for one kind of output. Keep each one narrow — one job, one shape. A skill that tries to cover every report produces output that fits none of them.

The signal that something should be a skill is simple: the second time you retype an instruction.

Step	What to do
Start from a repeat	Take an instruction you have already typed twice, verbatim
Write the output shape	Sections, order, length, and what to do when a number is not measured
Use it in a chat first	Check the shape against real workspace data before it goes on a cadence
Attach it to an automation	Once the shape holds, recurring runs produce comparable output week over week

Why this matters for AI visibility specifically

The output of this product is often text that will end up on your site — a rewritten section, a new FAQ answer, a page the agent proposes through a pull request. Text that does not sound like you is text you will not ship, and a suggestion you will not ship is worth nothing.

A brand profile is what turns “a plausible draft” into “a draft we can actually merge”. Combined with a skill that fixes the shape, a weekly run stops being a fresh negotiation about format and starts being something you read in thirty seconds.

A profile written as checkable rules

Constraints, not adjectives. Each line is something the agent can hold a draft up against and fail it.

Never call the product “a platform”. It is a product.

Always distinguish crawlers from models. They are separate systems.

Prices are USD, monthly, and always come from the pricing page.

Never claim a rank, a share-of-voice percentage or a sentiment score.

Buyers are in-house marketing leads at B2B SaaS companies, not agencies.

Skill “weekly-visibility-note”: 120 words max, three sections —

what moved and cleared the gate; what did not clear it; what was not measured.

Note the fourth rule. It is not just brand preference — it matches what the product can actually measure, so a draft that breaks it would also be a draft that is wrong.

What it does not do

Knowledge and skills cannot make the agent assert something the data does not support. This is where a brand profile stops being a wish list.

A profile saying “we lead the category” does not turn an unmeasured cell into a number.
A skill asking for a share-of-voice ranking cannot produce one. Attensira measures mention and citation rates against a shared denominator, so brand rates do not sum to 1 and there is no brand rank metric to report.
If a skill asks for a metric that does not exist, you get an honest gap in the output rather than an invented figure.
It does not change sampling. Depth stays at n=1 on Starter and n=3 on Growth and Business, and each Prompt × Model × country is still measured at most once per workspace-local calendar day.
It does not change mention rate, citation rate or source share. Those are measurement, not configuration.
Write skills around what exists: mention rate, citation rate, source share and average citation position.

Which plans include it

Knowledge & Skills is included on Growth ($349/mo) and Business ($699/mo). On Starter the workspace shows an upgrade state instead.

On Starter the workspace shows an upgrade state that names the plan requirement and does nothing else — it is not a broken or empty version of the feature, and nothing you configure elsewhere is lost by being on Starter. The agent still works in chats and automations; it simply carries no stored brand context.

Plan facts on this page are read from the same ladder that prices /pricing, so they cannot drift from what you are actually charged.

Read further

Documentation

Knowledge and skills — the brand profile, skills, and what they will not do

Documentation

Limitations — why there is no brand rank and no share-of-voice score

The agent

The profile and skills apply to every chat in the workspace.

Automations

Attach a skill so a weekly run comes out the same shape every week.

Pricing

Knowledge and skills is one of exactly two plan-gated capabilities.

Inbox

Where a draft that touches your site waits for an explicit yes.

FAQ

Questions people actually ask

Answers here match the product documentation. Where the product cannot do something, this says so.

Still deciding? Compare plans or talk to us.

A brand profile the agent checks its own drafts against

A brand profile

Named, reusable skills

Rules, not adjectives

It cannot invent a number

What it changes, and what it deliberately does not

The brand profile

Skills

Why this matters for AI visibility specifically

A profile written as checkable rules

What it does not do

Which plans include it

Read further

Questions people actually ask

Does a brand profile change my mention rate?

What should go in the brand profile?

What is the difference between the profile and a skill?

Can I ask a skill for a share-of-voice ranking?

Which plans include it?

If I downgrade, do I lose what I wrote?

What it changes, and what it deliberately does not

The brand profile

Skills

Why this matters for AI visibility specifically

A profile written as checkable rules

What it does not do

Which plans include it

Read further

Questions people actually ask

Does a brand profile change my mention rate?

What should go in the brand profile?

What is the difference between the profile and a skill?

Can I ask a skill for a share-of-voice ranking?

Which plans include it?

If I downgrade, do I lose what I wrote?