probeoBook a callPricing claims are stated without the conditions that change them.
High impactPer-employee pricing is given as a single figure while the page also describes plan minimums and add-ons. State what the figure includes and what moves it.
YMYL Accuracy
Content Intelligence
Know which pages still hold up.
What each page is trying to do, whether it still does it, and whether what it says holds up. Everything in Site Clarity underneath.
$3 per managed page, per month
Unlimited seats. Nothing to install. Every page you choose, read continuously.
Content
All questions
1How much does payroll software cost for a small team?example.com/pricing/→
2What does a payroll service provider actually do?example.com/guides/payroll-service-provider→
3How do I run payroll for the first time?example.com/guides/first-payroll→
4What payroll taxes does an employer pay?example.com/guides/employer-payroll-taxes→
5Can I switch payroll providers mid-year?example.com/guides/switching-providers→
6Do I need payroll software for one employee?example.com/guides/single-employee-payroll→
7What is a pay period?example.com/glossary/pay-period→
what you see
What every page is doing.
For each page, the question it answers, what it is for, where it sits in the journey, and the intents it carries. Its quality is graded against Google's rater guidelines (effort, originality, skill, accuracy, and experience, expertise, authority and trust), with the reason written next to each grade.
Claims that are wrong or overbroad are flagged, with what makes them so. The URL, title, H1, structure and calls to action are checked against the page's declared type. Then there are the topics it covers, its heading outline and reading level, and a short list of what to improve, ranked by impact.
Question analysis
The page walks visitors through choosing a payroll plan, compares the Basic, Plus and Premium tiers, shows what each includes, and repeatedly directs visitors to start a free trial.
Lifecycle stageSTARTING
The page addresses small teams setting up payroll for the first time, through plan comparison, setup steps, and the CTA 'Start Free Trial'.
Observed intents
EDUCATEObserved
The plans section explains tax filing, direct deposit, benefits, time tracking and compliance coverage for each tier.
EVALUATEObserved
The page enables plan evaluation through Basic, Plus and Premium pricing, feature comparisons, support details, and customer reviews.
DECIDEObserved
The page presents direct action CTAs such as 'Start Free Trial,' 'Get Started,' and 'Talk to Sales'.
Content domains
PAYROLL PROCESSINGObserved
The page focuses on running payroll, pay schedules, direct deposit, and the plan that fits a team's size.
TAXES FILINGObserved
The plan comparison covers federal, state and local payroll tax filing, and year-end W-2 and 1099 forms.
EMPLOYEE BENEFITSObserved
The Plus and Premium tiers describe health benefits, retirement plans and workers' compensation.
COMPLIANCEObserved
The FAQ discusses new-hire reporting and labor law posters as included services.
BUSINESS OPERATIONSObserved
The page addresses onboarding, time tracking, and integrations with accounting software.
the difference
All you get is TOO SHORT.
A crawler tells you a title is too short. It cannot tell you what the title should say, or too short for what.
It finds hundreds of pages where the title matches the H1, and cannot tell you which of the two should change, or whether matching is right on that page. Content Intelligence reads the page and decides, and shows you the reason.
Content question
How much does payroll software cost for a small team?
https://example.com/pricing
Service PageYMYLPricing
Page qualityLow+
A substantial pricing page with clear plan structure and support information. It is YMYL content, and the accuracy gate is triggered: several pricing and tax-filing statements are generalized or insufficiently qualified, so the rating cannot exceed Low. The organized structure supports Low+ rather than Lowest.
2 high 2 medium 1 low
- What needs to be fixed5
- Declared factual issues2
- What this page is forClear
- Main content quality4 areas
- Content alignment5 checks
- Question analysis3 intents
- Content snapshot1,412 words
- Impression rank
- #4
- Impressions
- 212,480
- Clicks
- 9,904
- CTR
- 4.7%
- Avg position
- 6.1
Improvement priorities
High
Tax filing coverage is asserted without naming its limits.
High impactFiling support differs by state and by plan. Say which states and which plans, or qualify the claim.
E-E-A-T
Medium
The page combines pricing, comparison, onboarding and support in one structure.
Medium impactSeparate the pricing decision from the onboarding explanation, and link the detail rather than carrying it here.
Quality of MC
Plan descriptions do not show feature-level differences.
Medium impactGive a comparison covering included services, add-ons, per-employee charges and limits.
Quality of MC
Low
The title-to-H1 relationship is only partially aligned.
Low impactThe title promises pricing for small teams; the H1 names the product. Keep the benefit and add the intent.
Title
Declared factual issues
"Every plan includes tax filing in all 50 states."
The statement is overbroad: filing support varies by plan and by state, and the page describes exceptions elsewhere without qualifying this sentence.
use cases
Where it earns its keep.
what it finds
Nine things about every page.
Archetype
Is this a guide, a service page, a tool, a hub, a policy page?
Intent
What question is this page answering, and is it answering more than one? Competing intents on a page are why it ranks for nothing.
Funnel position
Where this page sits in the journey.
Role drift
Has this page evolved away from what it was built to do? Pages drift when several teams edit them over years, because each edit was reasonable.
Canonical identity
Do the URL, the title and the H1 agree on what this page is? If they disagree, so does Google.
Brand voice
Alignment, and drift from it.
Question similarity
Across the whole site. Where several pages answer the same question and split their own authority.
Page quality
Graded against Google's own quality rater guidelines: effort, originality, accuracy, and experience, expertise, authority and trust. The reason is written next to every grade.
Factual claims
Statements that are wrong, overbroad or misleading, flagged with what makes them so.
why it is different
Each of these covers part of it. Most of what they call quality is keyword coverage.
The content tools you may already own, against the nine things Content Intelligence reads on every page. Every mark comes from the vendor’s own documentation.
| Archetype | Intent | Funnel position | Role drift | Canonical identity | Brand voice | Question similarity | Page quality | Factual claims | How it reads your pages | |
|---|---|---|---|---|---|---|---|---|---|---|
| probeo | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Every page, re-read monthly |
| MarketMuseIntent and funnel stage come from the keywords a page ranks for, not the page. Duplicate pages are pages ranking for the same topic, checked by hand. Content Score counts topic terms; its content quality analysis does not document how it judges. | Not offered or not documented | Partly | Partly | Not offered or not documented | Not offered or not documented | Not offered or not documented | Partly | Partly | Not offered or not documented | Site inventory updated about monthly; analysis on demand |
| ClearscopeOne target query per page, from Search Console. Content Grade measures term coverage and is re-scored monthly, which is its drift signal. A voice can be imitated when drafting, not scored. Overlap is found by filtering by hand. | Not offered or not documented | Partly | Not offered or not documented | Partly | Not offered or not documented | Partly | Partly | Partly | Not offered or not documented | 50 to 300 monitored pages by plan, rescanned monthly; drafts one at a time |
| SurferContent Score measures keywords, coverage and structure against the top results. Cannibalization is pages sharing Search Console rankings. Custom Voice guides writing; it does not score pages. The audit checks the keyword is in the title or H1, not that they agree. | Not offered or not documented | Partly | Not offered or not documented | Not offered or not documented | Partly | Partly | Partly | Partly | Not offered or not documented | One draft at a time; the site audit covers your top 100 pages |
| ConductorPage type, intent and journey stage are set on one draft at a time. The change history tracks page elements, not purpose. Cannibalization counts pages ranking for the same tracked keyword. Voice profiles guide writing and are not scored. | Partly | Partly | Partly | Partly | Not offered or not documented | Partly | Partly | Partly | Not offered or not documented | Crawls the site around the clock for technical changes; content scored one draft at a time |
| SiteimproveBrand Consistency scores pages against a voice profile, run on demand. Duplicate Content finds text overlap. Intent is per keyword; funnel stage is for planned content. Its content quality analysis (MarketMuse-based) does not document how it judges. | Not offered or not documented | Partly | Partly | Not offered or not documented | Partly | Yes | Partly | Partly | Not offered or not documented | Site crawl for text checks; brand and quality checks on demand |
| SemrushIntent is labelled per keyword. Tone of voice is a casual-to-formal scale on one draft. Cannibalization counts pages ranking for the same tracked keyword. Titles and H1s are checked separately, not against each other. Scores are term coverage and readability. | Not offered or not documented | Partly | Not offered or not documented | Not offered or not documented | Partly | Partly | Partly | Partly | Not offered or not documented | Content scored one draft at a time; rankings tracked daily |
probeo
Every page, re-read monthly
- Archetype: Yes
- Intent: Yes
- Funnel position: Yes
- Role drift: Yes
- Canonical identity: Yes
- Brand voice: Yes
- Question similarity: Yes
- Page quality: Yes
- Factual claims: Yes
MarketMuseIntent and funnel stage come from the keywords a page ranks for, not the page. Duplicate pages are pages ranking for the same topic, checked by hand. Content Score counts topic terms; its content quality analysis does not document how it judges.
Site inventory updated about monthly; analysis on demand
- Archetype: Not offered or not documented
- Intent: Partly
- Funnel position: Partly
- Role drift: Not offered or not documented
- Canonical identity: Not offered or not documented
- Brand voice: Not offered or not documented
- Question similarity: Partly
- Page quality: Partly
- Factual claims: Not offered or not documented
ClearscopeOne target query per page, from Search Console. Content Grade measures term coverage and is re-scored monthly, which is its drift signal. A voice can be imitated when drafting, not scored. Overlap is found by filtering by hand.
50 to 300 monitored pages by plan, rescanned monthly; drafts one at a time
- Archetype: Not offered or not documented
- Intent: Partly
- Funnel position: Not offered or not documented
- Role drift: Partly
- Canonical identity: Not offered or not documented
- Brand voice: Partly
- Question similarity: Partly
- Page quality: Partly
- Factual claims: Not offered or not documented
SurferContent Score measures keywords, coverage and structure against the top results. Cannibalization is pages sharing Search Console rankings. Custom Voice guides writing; it does not score pages. The audit checks the keyword is in the title or H1, not that they agree.
One draft at a time; the site audit covers your top 100 pages
- Archetype: Not offered or not documented
- Intent: Partly
- Funnel position: Not offered or not documented
- Role drift: Not offered or not documented
- Canonical identity: Partly
- Brand voice: Partly
- Question similarity: Partly
- Page quality: Partly
- Factual claims: Not offered or not documented
ConductorPage type, intent and journey stage are set on one draft at a time. The change history tracks page elements, not purpose. Cannibalization counts pages ranking for the same tracked keyword. Voice profiles guide writing and are not scored.
Crawls the site around the clock for technical changes; content scored one draft at a time
- Archetype: Partly
- Intent: Partly
- Funnel position: Partly
- Role drift: Partly
- Canonical identity: Not offered or not documented
- Brand voice: Partly
- Question similarity: Partly
- Page quality: Partly
- Factual claims: Not offered or not documented
SiteimproveBrand Consistency scores pages against a voice profile, run on demand. Duplicate Content finds text overlap. Intent is per keyword; funnel stage is for planned content. Its content quality analysis (MarketMuse-based) does not document how it judges.
Site crawl for text checks; brand and quality checks on demand
- Archetype: Not offered or not documented
- Intent: Partly
- Funnel position: Partly
- Role drift: Not offered or not documented
- Canonical identity: Partly
- Brand voice: Yes
- Question similarity: Partly
- Page quality: Partly
- Factual claims: Not offered or not documented
SemrushIntent is labelled per keyword. Tone of voice is a casual-to-formal scale on one draft. Cannibalization counts pages ranking for the same tracked keyword. Titles and H1s are checked separately, not against each other. Scores are term coverage and readability.
Content scored one draft at a time; rankings tracked daily
- Archetype: Not offered or not documented
- Intent: Partly
- Funnel position: Not offered or not documented
- Role drift: Not offered or not documented
- Canonical identity: Partly
- Brand voice: Partly
- Question similarity: Partly
- Page quality: Partly
- Factual claims: Not offered or not documented
✓ covers it◐ partly— not offered, or not documentedi what the marks leave out
From each vendor’s own documentation, September 2026.
why now
Search shows an answer, not a list.
8% vs 15%
When Google shows an AI summary, people click a result link in 8% of visits, against 15% without one.
2.5 billion
AI Overviews have over 2.5 billion monthly active users.
+120%
Pages cited in an AI Overview get 120% more organic clicks per impression than pages that are not.
Whether a page gets used depends on whether a machine can tell what it is for.
probeo measures AI Overview suppression per page, with a confidence level, rather than guessing at it. The measure is still marked experimental.
After one customer rewrote roughly 900 pages, the share of their impressions coming from AI Overviews roughly tripled over the next two months.
every finding has a reason
You can see why.
Every grade comes with the sentence that caused it. A page capped at Low for accuracy shows you the claim that capped it, and where on the page it sits.
Some findings only exist because two readings sit side by side. A page written for someone starting their first business can measure at a college reading level. Neither number is wrong alone.
Page quality
LowThe page has a beneficial purpose, substantial organized content, transparent plan pricing, and aligned sign-up CTAs. Because it is clear YMYL content and the accuracy assessment is low due to overbroad or insufficiently qualified pricing and tax claims, the YMYL accuracy gate limits the rating to Low.
What you can do with it.
Stop arguing about which pages to consolidate. The duplicate intent is measured.
Stop guessing why a page underperforms when its technical health is fine. The intent mismatch is visible.
Legitimately one of the most impressive things I've seen.
Erik T., Bizee
questions
What people ask.
Will probeo change our website?
Only if you want it to. It can rewrite pages and ship them, or hand everything it finds to your team.
How do you keep AI-written content accurate?
Research and writing are separate steps. The writer drafts only from facts that were researched and sourced, so it cannot invent one. Every page is then graded against Google's quality rater guidelines before it ships, and you decide what goes live.
Do you track visibility in ChatGPT and other AI tools?
probeo measures AI Overview suppression per page, marked experimental, and separates AI assistants in your referral traffic. Checking how AI tools answer your questions is not available yet.