Attensira for SEO teams

Extend your Google work to the AI answers, with the same rigour.

For an SEO team, Attensira is software that does the AI-search visibility work: it reads the AI answers beside your Search Console data, shows which sources each engine leans on, opens the on-page and technical fixes as pull requests, and re-reads every change at +7, +14 and +30 days with its sample size.

7 days free · card requiredHow the trial works

Before and afterEvery verdict carries its n

FAQ structured data and a rewritten intro on /pricing

+7 daysNo changen=42
+14 daysToo close to calln=42
+30 daysWaitingnot read yet

Example · fictional business, not a customer

What does the crew see on our own site?

We run Attensira on attensira.com. Sites we own, not customers. A baseline, not a result. These are the answers before the crew's work on our own site, and they include where we lose.

How we measure

attensira.com · ChatGPT and Perplexity24 of 31 days sampled

239of 4,518 answers linked to attensira.com

of 4,518 answers named attensira.com: 197
of 100,248 citations pointed at our own sites: 1,588

Window 2026-09-04 to 2026-10-04. Other engines were not read here: absent, not zero.

attensira.com's own workspace, read over the Attensira MCP · 4 Oct 2026 · n=4,518 answers

How does it fit an SEO team's work?

Your method, extended from the results page to the answer.

Step 1:
Read both result pages
Search Console clicks, impressions and positions beside the AI answers for the same pages, so one page's two results sit together.⁠
Step 2:
Ship the fixes as PRs
Structured data, meta tags, alt text, internal links and copy changes on a fresh branch. Hygiene can run on its own in Act; copy always waits.⁠
Step 3:
Hold it to a verdict
Each change is re-read at +7, +14 and +30 days. No change and too close to call are verdicts, and every one carries its n.⁠

What does the first week look like?

The trial week comes with $30 of credits (3,000 credits): it reads every tracked prompt on each engine, shows the deciding sources, and the agent drafts real fixes you review and approve. Outreach, automations and top-ups start after the trial.

The trial, day by day

Before the card
Your domain, then a free scan
You give your website. The agents read it and propose your business, competitors and starter prompts for you to correct, and a free scan shows how often AI answers name you, usually before you are asked for a card.⁠
Today
Card checked, nothing charged
Checkout states the charge before it takes the card. The Starter plan fee is not charged until the trial ends.⁠
Day 1
First answers in your workspace
The moment the trial starts, every prompt you track is read on every engine your plan includes, and you watch the rows fill in. The first report lands once half the reads are in, or after twenty minutes. After that, every prompt is read once a day.⁠
Days 1 to 7
The answer grid
One row per prompt, one cell per engine: named and linked, named, lost to a rival, or absent. Each cell is read from the answers that came back, and says how many there were.⁠
Days 1 to 7
The sources behind each answer
For every prompt on every engine, the pages the answers leaned on, ranked by weight rather than raw citations, and whose each one is: yours, a competitor's, or someone else's.⁠
Days 1 to 7
What is missing, and why
A read that failed says so instead of counting as a zero. A cell with nothing to read is drawn as not measured, never as a loss, and says why: not read yet, every read failed, or no answers came back.⁠
Days 1 to 7
The agent drafts real fixes
The trial includes $30 of credits (3,000 credits), and the agent spends them on real work for what the readings found: a rewrite, a page fix, an answer passage. Each draft waits in your Inbox with what it found and why. Approve what you like. Nothing ships without your yes, unless you let a lane ship on its own.⁠
Day 7
The trial ends, and the plan starts
The Starter plan fee, $129 a month, is charged on day 8 unless you cancelled first. Then the plan's own credits start, and with them outreach, automations and top-ups. Cancel in the app before the trial ends and the plan fee is never charged. No call to book, no email to send.⁠

What waits for an SEO lead's yes?

Nothing ships without your yes, unless you let a lane ship on its own.

Three modes: Ask, Draft, Act. Draft is the default, so everything that would leave the workspace waits for you.⁠
Claims wait in every mode: a number, a price, a product or competitor name, a promise.⁠
An approval the crew asks for waits up to 7 days. If you never answer, the run goes on without doing that thing.⁠

InboxApprovals

Approval mode

The crew does the work inside the workspace and files the rest for your yes. Every workspace starts here.

Always waits for you

Rewritten heading words or a title tag⁠Waits for your yes
Merging any pull request, unless you turn on auto-merge⁠Waits for your yes

May run on its own, in Act

FAQ and JSON-LD blocks⁠Waits for your yes
Internal links between your own pages⁠Waits for your yes

4 of 4 waiting for you in Draft

Example · how each mode treats this work

Where do the fixes land?

GitHub: Pull requests on a fresh branch, inside the directories you said hold your pages, plus llms.txt, robots.txt and sitemaps at the root.⁠Never code blocks, source files, CI, lockfiles or dotfiles, even with a yes.
Your Inbox: The diff to review, and Merge pinned to the commit you reviewed.⁠Never merged unless you press Merge or turned on auto-merge for safe changes.
Page changes: Every change to a tracked page is seen and measured, whoever made it.⁠Never credited as the cause of a movement.

How else does an SEO team cover AI answers?

The real competitor is the week nobody has.

Four ways to get the work done, and what each leaves you
Option	What it looks like
Do it yourself	Manual prompt checks in five assistants, pasted into a spreadsheet each Friday.
Hire an agency	A GEO add-on to the SEO retainer, with the tickets still landing in your backlog.
Buy a dashboard	Visibility scores with no fixes attached and no sample size beside them.
The crew	Fixes arrive as reviewable pull requests and come back with a verdict and its n.

What does it cost?

A plan fee for monitoring, and credits for the work. Every number here comes from the price list the app itself uses.

The Starter trial is 7 days with $30 of credits (3,000 credits): you watch your readings come in and the agent draft real fixes, and you approve what you like. Outreach, automations and top-ups start after the trial. Card required; cancel in the app before it ends and nothing is charged. Upgrade in the app under Settings → Billing.⁠
Starter samples each prompt once a reading (n=1) on 4 engines: ChatGPT, Perplexity, Claude and Google AI. Growth and Business sample 3 times and add Gemini, so a real movement proves out sooner.⁠
Credits pay for the crew's work: the more credits you buy, the more work it can do, inside a monthly budget you set.⁠
We do the work and charge only the credits it used, never a promise of outcomes or of a set volume.⁠

Compare plans

Growth

$349/mo

Three samples a reading let real moves clear the significance gate sooner, and Growth adds Gemini and AI traffic.

Prompts tracked: 150
Engines: 5: ChatGPT, Perplexity, Claude, Gemini, Google AI
Samples a reading: n=3
Credits included a month: 7,500
History: 12 months

Start on Starter, upgrade in the app

7 days free · card required · $30 of credits (3,000 credits) · on Starter

What if it doesn't work, and what can it touch?

If it is not for you

Cancel in the app before the trial ends and nothing is charged. The trial's drafts wait for you: Nothing ships without your yes, unless you let a lane ship on its own.⁠
A pull request is a proposal: close it and the change is gone. A CMS change is a draft until you publish it.⁠
Your receipt lists everything the crew did, and says plainly when nothing is proven yet.⁠

What it can touch

Only the repositories you pick, through the GitHub App. A fresh branch per change, never your default branch, never your CI.⁠
In your CMS, drafts in your own editor. A new article is never published for you.⁠
Pitches as drafts in your own mailbox. Nothing is sent from an Attensira domain.⁠

What it never touches

Code, source files, CI configuration, lockfiles and dotfiles, in any mode, even with a yes.⁠
Mail from your domain, money or anything irreversible, without an explicit yes.⁠
Its own approvals: an agent never approves its own work, and only a person raises what it may do alone.⁠

Security

What do SEO teams ask before starting?

Will it fight my existing SEO?

It edits only inside the directories you say hold your pages, and any copy or heading change waits for your yes. Hygiene changes run on their own only if you choose Act. In the docs

What exactly can it change without asking?

In Act mode only: structured data, meta descriptions, alt text, internal links and heading levels, and only when every line it touches is on that list. In the docs

How many changes can it make on its own?

Act mode is capped: by default two unattended changes a week across the workspace and one per page. A change you approve yourself is not rationed. In the docs

How does it decide which sources matter?

By weight, not raw citation counts, for each prompt on each engine, classed as yours, a competitor's or someone else's. In the docs

Why does a change read as no change?

Models are not deterministic. At one sample a reading most movements do not clear the 95% gate; at three samples real movements clear it sooner. In the docs

Does it see changes we made ourselves?

Yes. A change to a tracked page is seen and measured whoever made it, and never credited as the cause. In the docs

Where should you go next?

Bring the AI answers into your SEO work.

7 days free · card requiredHow the trial works

Extend your Google work to the AI answers, with the same rigour.

What does the crew see on our own site?

How does it fit an SEO team's work?

Read both result pages

Ship the fixes as PRs

Hold it to a verdict

What does the first week look like?

Your domain, then a free scan

Card checked, nothing charged

First answers in your workspace

The answer grid

The sources behind each answer

What is missing, and why

The agent drafts real fixes

The trial ends, and the plan starts