What this screen is
Once your profile is published, two kinds of activity accumulate: visitors ask your profile’s AI questions, and hiring managers run fit assessments against you. AI Activity (available on every tier) is the raw log of both. Insights is the Pro layer on top: the patterns in that activity, read for you.
The screen’s heading reads "What the AI keeps concluding about you". Nothing on it is new data: it is your AI Activity log, read for what repeats, together with your own practice debriefs and your publishing gates.
Three facts about the screen:
- Who sees the content: Pro. Insights is part of Pro, not a separate purchase, and Pro that you received as a complimentary plan or earned through referrals includes it in full. The page is visible on Free with an upgrade prompt.
- When it appears: the Insights entry sits next to AI Activity in the sidebar only when the screen is switched on for your account by the Profiled team. When it is not switched on, the entry is absent and the page address shows a not-found page. Your User Dashboard at a glance: the menu, tiers, flags and banners explains why your menu might not match someone else’s.
- What it is made of: every item on the page is a count, a label, a percentage or a fixed sentence. None of it is AI-written prose about you. When it says a requirement came up in 6 of 9 assessments, that is arithmetic on your own assessments, and the interpretation is left to you.
What you see on Free
On Free you see the three activity counts (once your profile has any activity) and, below them, a card labeled "Pro feature" and headed "See what the AI keeps concluding about you". It has two buttons:
- Upgrade to Pro opens the pricing page. For what Pro costs, see pricing.
- View AI Activity opens the raw log.
None of the sections described below is shown on Free. The counts are shown because a count is a fact about how much activity your profile has, not a conclusion drawn from it.
The empty state
Insights needs activity to read. A freshly published profile shows "No activity to read yet" — that’s the expected state, not a fault. Share your profile link; the page fills as people engage.
The empty state offers two buttons:
| Button | Where it goes |
|---|---|
| "Strengthen your profile" | The Profile Builder |
| "Share your profile link" | The Refer & Earn screen. It does not copy your profile address |
Below the empty state, Free shows the upgrade card. Pro shows the Profile readiness section, because your publishing gates do not depend on anyone having visited.
The empty state goes away as soon as any one of the three counts is above zero. A practice session of your own with a saved debrief is enough.
The activity counts
Three cards, shown on both plans once your profile has any activity.
| Card | What it counts |
|---|---|
| "Fit assessments run" | Fit assessments visitors ran against your profile. Checks you run yourself on Check a Job Fit are left out of this count and of every pattern below; an assessment you run on your own public profile page counts like a visitor’s |
| "Visitor questions asked" | Questions put to your public profile’s AI, including clicks on the suggested questions shown under the chat |
| "Practice sessions completed" | Your own practice interviews that have a saved debrief |
The third card is the difference from AI Activity: Insights also reads your own practice sessions. The voice interview you may have used to build your profile is not a practice session and is not counted.
How your assessments land
Fit assessments — how many have run, and how your assessments land: the mix of verdicts your profile is receiving. The line under the heading gives the total, for example "Across 9 assessments run against your profile."
There is one bar for each verdict you have received, with its count and percentage: Strong fit, Worth a conversation, Probably not your person. A verdict you have never received is left out.
The verdict mix is honest by design — including honestly negative. A Probably not your person verdict against a poorly-matched job description is the product working as intended, not your profile failing.
Two mixes are worth a second look. Mostly Probably not your person usually means the wrong roles are finding you: check "Target titles" on the Basics tab, "Looking for" on the Narrative tab, and where you share your link. Almost all Strong fit can mean your gaps are underspecified: check the Gaps tab and the "Still developing" column on the Skills tab.
The trend line. One sentence under the bars compares the earlier half of your assessments with the more recent half.
| Direction | The line reads |
|---|---|
| Improving | "Your recent assessments are landing better than your earlier ones." |
| Declining | "Your recent assessments are landing worse than your earlier ones." |
| Steady | "Your verdicts have held steady over time." |
| Too few to tell | "Not enough assessments yet to show a direction — this appears once you have at least four." |
With two or three assessments, one unusual job description would swing the result, so the page says it does not know yet. A declining line is often about the roles: check the job descriptions in AI Activity before you change your profile.
How sure the AI was. Under the trend line, a row labeled "How sure the AI was" shows how many assessments fell into each band: "High confidence", "Moderate confidence", "Low confidence". A band with no assessments is left out. Read it as a caveat on the mix above.
If your profile has visitor questions or practice sessions but no fit assessments yet, this section and the two after it are replaced by one section headed "Fit assessments" with the line "Nobody has run a fit assessment against your profile yet."
What keeps coming up as a gap
What keeps coming up as a gap — the recurring themes on the gap side of your assessments. If three different job descriptions all surface the same missing skill, that’s signal about your profile — or about the roles being tested against it.
The line under the heading reads "Requirements flagged in 2 or more of your assessments." Each row is a requirement phrase taken from the job descriptions, with a count such as "4 of 9": four of your nine assessments raised it. An assessment counts once for a theme, however many of its gaps name it. A requirement that came up once is not listed.
Two requirements count as the same theme only when their wording matches, apart from small differences such as capital letters and punctuation. The same requirement phrased two ways is counted as two, so this list can be shorter than you expect. With nothing recurring, the section opens "No requirement has come up more than once yet."
Recurring gaps are the actionable part. A gap theme you can genuinely close belongs in your Skills or Experience tabs; a gap theme you don’t intend to close might belong in your Gaps tab, stated honestly — an acknowledged gap reads better in an assessment than an unexplained absence.
That gives three honest responses:
| If the gap is | Do this | Where |
|---|---|---|
| Something you have but did not write down | Add it, with evidence | Tab 4 — Skills · Tab 3 — Experience |
| Something you lack | Declare it | Tab 5 — Gaps · Skills tab, "Still developing" column |
| Something irrelevant to the roles you want | Leave your skills alone and fix your targeting | Basics tab, "Target titles" · Narrative tab, "Not looking for" |
An example using Elena Marchetti, a demo profile. The themes and counts are made up to show the three responses; they are example content, not a readout of her profile.
| Gap theme | Count | An honest response |
|---|---|---|
| People management experience | 5 of 9 | Declare it on the Gaps tab and add it to "Not looking for", so roles that need a people manager stop arriving |
| Pricing experiments | 3 of 9 | If she has done the work and left it off, add it with evidence. If she has not, declare it |
| Kubernetes | 2 of 9 | Already on the Skills tab as "Still developing". Nothing to change |
What keeps coming up as a strength
What keeps coming up as a strength — the same, on the strength side. Worth reading before an interview: it’s what the assessments are consistently saying for you.
The line under the heading reads "Areas credited in 2 or more of your assessments." Rows carry the same "4 of 9" style of count. With nothing recurring, the section reads "No strength has been credited more than once yet."
Use it to check your positioning. If a strength that keeps being credited is buried low in your Narrative tab, move it up, and make sure your elevator pitch leads with it.
What visitors ask your AI
Visitor questions — the themes in what people ask your profile’s AI.
The line under the heading gives the number of typed questions, for example "7 questions typed by visitors, grouped by topic." Before anyone has typed one it reads "No visitor has typed a question yet."
Each topic shows a label, a count such as "asked 3×", up to three of the questions quoted word for word, and a line such as "Answered by your Experience section." The tab name in that line is a link that opens the tab in the Profile Builder. A topic is listed even if it was asked once.
There are ten named topics, such as "Why they left a role", "Gaps and weaknesses", "Salary and rate" and "A specific tool or technology". A question that fits none of them goes under "Other questions", which points to the FAQ tab.
Typed and suggested questions are counted apart. Visitors see suggested questions under your chat. A click on one of those is Profiled’s suggestion, not the visitor’s own curiosity, so those clicks are left out of the topics and reported on a separate line at the bottom of the section, for example: "3 further clicks came from the suggested questions we show under your chat, so they are counted separately — they tell you what we prompted, not what a visitor wanted to know." Base your decisions on the typed questions.
To act on a topic:
- Read the quoted questions, not only the topic label.
- Follow the tab link and check whether that tab answers the question.
- Give anything asked three times or more an entry on Tab 7 — FAQ.
- Check that the answer is in a public field. The visitor chat is given your public fields only. What does the AI chat on my public profile know about me? lists them.
What keeps coming up in practice
This section reads your practice debriefs, if you use voice practice. It looks at your six most recent practice sessions with a saved debrief and lists each topic from their "Weak" columns that appeared in two or more of them, with a count such as "2 of your last 4". The line under the heading reads, for example, "Across your last 4 practice sessions." The debrief’s "Question by question" labels are not counted, and neither are practice-only questions, which are never graded.
Only the topic is shown. What the debrief said about how you answered stays on the debrief, on the Practice screen.
Before your first session the line reads "You have not completed a practice interview yet." and the section offers a "Start a practice session" link. With sessions but nothing repeating, it reads "No weakness has appeared in more than one of your recent sessions yet."
Tip: compare this list with your recurring gaps. A topic in both lists is one that job descriptions keep asking for and that you cannot yet back up out loud, so it comes first. A topic here that is not in the gap list is a delivery problem: rehearse it; your profile does not need rewriting.
Profile readiness
The last section, and on Pro the only one shown when there is no activity.
The line under the heading gives your score and the state of the gates, for example "Quality score 58 / 100. 2 gates still open." When nothing is open it ends "Every quality gate is met."
The five publishing gates are listed by name: "Experience depth", "Skills breadth", "Professional narrative", "Acknowledged gaps" and "Your name". Each has a colored dot, a one-line explanation, and, while it is unmet, a "blocks publishing" tag and a link such as "Edit Experience" that opens the tab that closes it. For "Your name" the link reads "Edit Basics".
These are the same gates the Profile Builder shows next to the Publish button. Why won’t the Publish button enable? covers each gate, and How the quality score works covers the score.
A monthly review
About fifteen minutes, in this order, which is not the order of the page.
- Profile readiness. An open gate outranks everything else, because an unpublished profile collects no activity.
- What keeps coming up as a gap. Take the top one or two and decide: add, declare, or fix your targeting.
- What keeps coming up in practice. Anything that is also in the gap list moves to the front.
- What visitors ask your AI. Anything asked three times or more becomes an FAQ. Follow the tab links.
- How your assessments land. Check the mix, and check whether the trend is explained by the roles.
- What keeps coming up as a strength. Make sure your top recurring strength is in the first line of your elevator pitch.
- Make the edits. A review that changes nothing in your profile was only reading.
Insights vs AI Activity
| AI Activity | Insights | |
|---|---|---|
| Tier | Every tier | Pro |
| What it shows | Every visitor question and your most recent fit assessments, individually | The patterns across them |
| Where | Sidebar → AI Activity | Sidebar → Insights |
If you’re on Free and want to see what’s being asked, AI Activity already shows you the items themselves, one at a time (every question, and your 20 most recent assessments) — Insights adds the rollup.
Finishing
- For each recurring gap you chose one response: add, declare, or fix your targeting
- Anything visitors asked three times or more has an FAQ entry
- Your top recurring strength is in your elevator pitch
- No gate under Profile readiness is still open
- You made at least one profile edit as a result
Related
- AI Activity: what visitors asked and how the AI answered
- Voice interviews and practice (Pro)
- Why won’t the Publish button enable?
Last verified: 2026-10-21