Skip to content

SEO automation with a human in the loop: the weekly work, run for you.

Describe the job once. A workflow runs it on schedule: pull the data, research the gap, draft in your brand's voice, pause for a person to approve, then publish.

First growth signal in ~2 minutes · no pixel to install

A content-refresh workflow on its Monday run: Search Console decay in, five refresh drafts out for an editor to review.

Monday morning

Someone exports Search Console every Monday.Or a workflow does.

Sorting the pages losing clicks, asking each engine who it cites instead, drafting the refreshes: the same job every week, done only when someone remembers. Describe it once and it runs on its own clock.

One run, step by step

Monday's job, done before you're in.

Follow one run of the content-refresh workflow from its trigger to the drafts it leaves for your team, with what each step cost.

05 · The team is told. The drafts wait for them.

Workflow · Content refresh from GSC decay

103 tasks this run

Every MondayTrigger · weekly, 09:00
run #14
Pages losing clicksGSC metrics · 90 days
5 slipping
× 5 pagesWho AI cites insteadAI citation check · 6 engines
0 of 6 cite
× 5 pagesResearch the gapPerplexity research
14 sources
× 5 pagesDraft each refreshLLM call · brand voice
5 drafts
Tell the teamNotify · Slack webhook
#content

5 refresh drafts · waiting for your team

next run · Mon 09:00

How it works

Describe it. Run it. Hand it over.

  1. 01 · Describe

    Describe the job

    Pick a trigger, such as a schedule, a webhook, a manual run or an API call, then chain the steps the job needs: Search Console metrics, an AI citation check, research and an LLM draft.

  2. 02 · Run

    Let it run on its clock

    The workflow runs unattended. Loops work through each page, branches handle the conditions, and each step's result is logged in the run so you can see what it did and how long it took.

  3. 03 · Hand over

    Leave the finished work for your team

    A run ends by notifying your team, staging rows in a grid or pushing to your CMS. This one leaves five refresh drafts and a note in #content, and waits for a person to take them from there.

    Run #14 · left for your team
    • /guides/pci-compliance-checklistdraft ready
    • /blog/chargeback-rate-benchmarksdraft ready
    • /guides/payment-orchestrationdraft ready
    • /blog/3ds2-explaineddraft ready
    • /blog/payment-routing-rulesdraft ready
    • #content · Slack webhooknotified

    Nothing went live: the drafts wait for your editor.

Building blocks

Steps that already know your data.

Start a run on a schedule, from a webhook or the API, or by hand before a planning meeting. Then chain steps that read your own numbers next to the open web, with loops, branches and filters to pass on only what the next step needs.

  • Search Console metrics and AI citation checks across six engines
  • Pages from your sitemap, web search, scraping and Perplexity research
  • End by notifying your team, staging grid rows or pushing to your CMS

schedule · webhook · manual · API

Step library
  • GSC metrics · pages and queries
  • AI citation check · 6 engines
  • Get pages from sitemap
  • Web search · scrape · Perplexity research
  • Loop · branch · filter
  • Notify · CMS publish · stage rows in a grid

Weekly, every two weeks or monthly, or on a webhook.

Off your plate

~30h
saved a week

Across the work Mentionova's agents take on.

4
ways to start a run

Schedule, webhook, manual or API.

6
engines checked per page

ChatGPT, Perplexity, Claude, Gemini, Google AI and Reddit.

≈103
tasks for this run

Five pages: GSC 3, then per page citation check 10, research 5, draft 5.

The AI-visibility workflows have been a goldmine for booked calls and closed deals. Up and up and to the right — the only direction it goes!
Mark RegalSaaS Founder

SEO Automation · FAQ

Questions, answered.

SEO automation is software that runs recurring search work on its own: pulling performance data, finding the pages that need attention, researching, drafting and publishing on a schedule instead of by hand. In Mentionova each job is a workflow, a trigger followed by a chain of steps and an output. Because answer engines now decide which pages get quoted, the steps include AI citation checks across ChatGPT, Perplexity, Claude, Gemini, Google AI and Reddit alongside your Search Console data. A human review step keeps a person in charge of anything that publishes.

A human-in-the-loop workflow runs the repetitive steps automatically and stops at the decision points. In Mentionova that stop is a human review step: the run pulls data, researches and drafts, then waits until a person approves, edits or rejects the work. Approved work continues to publishing or notifications; rejected work stays a draft. You get automation on the busywork and a person's judgment on what reaches your site.

The best candidates are weekly jobs with the same shape every time. Refreshing pages that lose clicks in Search Console, checking which sources the AI answers cite instead of you, drafting briefs for a list of topics and posting a summary for the team all fit. Work that needs a judgment call, like choosing positioning or approving what publishes, stays with a person through the review step.

An SEO automation tool runs recurring search work without someone clicking through it each week. Mentionova's workflows pull Search Console data, check which sources ChatGPT, Perplexity, Claude, Gemini, Google AI and Reddit cite, research the gap, draft in your brand's voice and publish to your CMS, pausing for a person's approval before anything goes live.

No. Put a human review step before the publish step and the run stops there until a person approves, edits or rejects the draft. Approved work continues to your CMS or to your team's channel; rejected work stays a draft. Nothing reaches your site on its own.

Four kinds of trigger: a schedule, such as every Monday at 9:00; a webhook from another tool; a manual run when you need results now; or a call from the API. Each one starts the same chain of steps, so a single workflow can run on a clock and on demand.

Workflows have built-in steps for Search Console metrics, AI citation checks across six engines, competitor overlap and pages from your sitemap. They can also run web searches, scrape pages, search Reddit, run Perplexity research and call an LLM, so one run can combine your own data with fresh research.

Yes. A workflow is a sequence of steps on a trigger, which suits recurring jobs. A grid is a table where each column does work across many rows at once, which suits bulk drafting. A workflow can stage its results as rows in a grid, and a grid run can be put on a schedule.

Most of the saving comes from the weekly jobs nobody wants: exports, sorting, first drafts and status updates. Across the work Mentionova's agents take on, teams get back about 30 hours a week. After you connect your domain, the first growth signal lands in about two minutes.

Free AI visibility report

Hand off theweekly work.

Connect your domain and Search Console, describe the job, and let the first run land before next Monday. Nothing publishes without your approval.