SEO automation software for the work nobody should do by hand

You can automate anything that is the same every time: collecting data, checking rules, spotting changes and measuring results. Judgement calls, such as what your brand should say or which bet to make this quarter, should stay with people.

The mistake most teams make is automating output before deciding what is worth doing. A tool that publishes a post a day will happily publish thirty posts nobody searches for. LogNorm automates the deciding first: every finding is ranked against every other piece of work before anyone writes a word. Our post on SEO automation with AI agents covers where to draw the line.

In practice, LogNorm automates the research and the ranking, and your AI agent picks up the hands-on fixes and drafts. Approval stays with your team. You can start with LogNorm alone and add agents later.

LogNorm runs the research you schedule and the rest of the growth loop without anyone clicking. Each website keeps its own schedules.

Scheduled runs need a paid plan or an active trial. If a run cannot start, for example because credits ran out, LogNorm records why, shows it in red and tries again after one full interval instead of retrying every 15 minutes. Details are in Automation and schedules.

LogNorm ranks moves against each other instead of scoring each one alone. It asks head-to-head questions, such as whether the team should do A or B first this week, weighing upside, odds of working, effort and urgency for your business.

Every pair is asked both ways round to cancel order bias, and the answers are fitted into one ranked list. New moves are slotted in without reshuffling everything, and your own Rank higher and Rank lower calls are kept on every pass. When two analyses find the same opportunity, LogNorm merges them and raises confidence instead of creating a duplicate. See Moves and ranking.

The hands-on work: fixing audit findings in your codebase, researching and drafting articles, building topic plans and running growth reviews. Your own Claude Code, Codex or Cursor agent does it, connected to LogNorm over MCP.

• Fix moves: the agent changes templates, metadata, robots.txt or sitemaps and asks LogNorm to re-check the live pages.
• Content moves: it researches a brief from the ranking pages, writes the draft and sends it for review.
• Keyword work: it targets or skips suggestions and builds hub-and-spoke topic plans.
• Reviews: it runs a growth review and posts findings with the moves it created.

The agent's thinking and writing cost no credits. See the AI SEO agent page for the fix loop.

Automate the research and the checks, not the judgement. LogNorm's content pipeline does the slow, checkable parts on its own and hands the result to a person before anything goes live.

• Briefs pull the live Google results with People Also Ask, read the top pages, and collect your Search Console queries and Company Brain passages.
• A fact is cited only when LogNorm opened the source, found the figure on that page and rated the site as original research, an official source or an established publication.
• Drafts review themselves for stock phrases, generic sections, claims that contradict your Company Brain and missed must-cover items, and rewrite what falls short once.
• Every link in the draft is opened; dead links, competitors and weak sources are removed.
• Quality checks score title and description length, keyword placement, heading structure, readability, internal links, schema and claims.

Failing checks don't block publishing, but reviewers see them. The editor and its scorecard are covered in Content studio.

Approving and publishing content, setting the order of the week and deciding which keywords you go after. LogNorm keeps those with people by design.

• Publishing: agents send drafts for review; a person approves and publishes.
• Keyword decisions: LogNorm skips a keyword only on a confident call and asks you about the rest in Needs your call. It never changes a decision a person made.
• Weekly plan: LogNorm recommends what to pick and never stops you planning more or less.
• Retries: a failed CMS publish is never retried automatically, because a blind retry could create a duplicate post.

Every shipped move is measured against its own Search Console baseline at 28 and 90 days. A move wins when clicks or impressions rise 20% or more, a new page reaches 10 clicks or 100 impressions, or average position improves by 3 places.

The Results page shows moves shipped, your win rate, and win rate by move type, so you learn which kinds of work pay off for your site. Pages that lose 30% or more of their clicks come back as refresh moves.