Find Answers Instantly.

Questions are answered from the same membership records the directory uses, under stricter rules: individuals appear only in professional lists (the same fields the directory shows), demographics are aggregate-only, and money never appears. The AI component itself never receives member records — it only translates the question into a controlled search plan.

Almost always one of four protections: contact details of named people (use the Member Directory instead), individual demographics or finances (never disclosed), lists filtered by a demographic (counts yes, names no), or mentor pairings (private). The refusal message states the reason; rephrasing toward a count usually gets the insight you wanted.

Your society masks very small counts so an ultra-specific question can't identify one person. The real number exists in the office's own reports; Ralph deliberately blurs it below the threshold.

Lists are capped to stay useful and respectful. Add a condition — a state, a subspecialty, a research method — and the right people surface.

Rephrase more concretely (the starter chips show the patterns that work), and tell the office the exact wording you used — misunderstood questions are reviewed and Ralph is taught the missing vocabulary. It genuinely improves.

No — Ralph answers questions; it never contacts anyone and never exports. Click through to a member card and email a colleague yourself, or bring legitimate mailing needs to the office.

Both, in that order: Ralph is the quick scan (“willing mentors in my field”), Mentor IntelliMatch (Chapter 9) is the real machinery — registration, AI-ranked pairings, and requests. Ask Ralph to look; use IntelliMatch to actually match.

Other members never see them. Usage is logged like the rest of the portal, and questions Ralph couldn't answer may be reviewed by the office — that's how it learns. Ask freely; just don't paste anything confidential into the box.

Rarely — and the page degrades gracefully: common question shapes are answered by a built-in fallback, so “how many members in Texas” works even on a bad day. Complex phrasings may need to wait for the AI to return.

Where your society uses the applications archive, yes — winners, applicants, and volumes by program and year, including historical years. Names and outcomes only; nothing from inside application forms.

You're in the pool: mentees see you in the mentor directory and Ralph may rank you for fitting mentees. Add a blurb on the Be-a-mentor tab (what you can actually help with) and keep your Professional-tab selections current — that's what the ranking reads. Requests, when they come, arrive by email and wait on your My-matches tab.

Flip Willing-to-mentor to No (or Maybe) — you disappear from the directory and the ranking immediately. Existing matches are untouched; the switch only controls new visibility. Flip back when a seat opens.

It reads what you wrote in 'what I'm looking for' plus your research profile, and compares against each registered mentor's blurb and profile — then ranks and explains each suggestion. It can only choose among mentors actually registered and willing; it never invents anyone. Sparse results usually mean a sparse profile — yours or the pool's.

The request is waiting on the mentor, who was emailed. You can withdraw it on My matches and request someone else — or do both: withdraw, then ask Ralph for fresh matches. The office can also gently nudge chronically silent mentors; tell them.

No — deliberately. They see your name and what you're looking for; contact details are exchanged only when they accept. A decline means nobody's inbox changed hands.

Nobody knows. Declines are between you and the mentor, carry no stated reason, and are usually about capacity, not you. Your registration stays active — ask Ralph for more matches; the second request is usually the one that lands.

Yes and yes — the platform doesn't ration relationships; mentors ration themselves with the willingness switch and their accept decisions.

The relationship lives between the two of you — end it the human way, by talking. The platform's record of the match is history, not a contract; the office can tidy records if ever needed.

You, your mentor, and — in aggregate only — the office's program statistics. Ralph refuses who-mentors-whom questions by design.

Sharpen the inputs: rewrite 'what I'm looking for' concretely (goals, stage, topic), and check your Professional tab selections. Then rerun — the ranking is only as good as what it reads. Persistent oddities are worth reporting with the exact text you used.

The phrases are composed from your research profile — generic in, generic out. Add or sharpen selections on My Profile → Professional (aim for three to six per panel), revisit the page, and the chips regenerate. The '(from your selections)' tag means the AI wasn't reachable and the chips are your raw labels — try again later for composed phrases.

Yes — every search queries Grants.gov and NIH RePORTER live at the moment you run it, and the funder list is the office's current version. Nothing is cached from last month. That's also why the answer to 'is this deadline right?' is always: open the source link and confirm there.

Occasional and transient — public federal services have moods. The card says so honestly; try again in a moment. The other sources still return.

If your search phrase carries no child-health signal, it quietly adds pediatric context so results skew to your world; searches already about children pass through untouched. Turn it off to search a topic outside pediatrics.

No — it finds and links. Eligibility, deadlines, and applications live at the source (that's the footer's caveat, and it's policy). Think of the page as the world's fastest research assistant, not a grants office.

Suggest it to the office — the list is curated by them precisely so it grows from member knowledge. Include the funder's grant-application page link if you have it.

Intelligence: funded projects show which study sections and institutes fund your area, at what scale, and who your funded peers are — the raw material of a competitive application. Many members find the RePORTER pile more valuable than the open-opportunities pile.

Usage is logged like the rest of the portal (which features are used, how much); other members never see your searches. The external services receive only the search phrases — never your identity or profile.

The manual search works immediately for any topic. But ten minutes on My Profile buys you the composed searches and the one-button match — worth it.

Mentoring is 'help me develop' — guidance, career, growth (Chapter 9). IntelliConnect is 'let's accomplish something together' — projects, papers, studies. Many members use both; the profiles are separate on purpose.

ON: you appear in the default open-to-collaboration listing and carry the badge. OFF: searches can still find you by your facets and text, but you're not advertised as looking. It's a visibility dial, not a wall.

Check the chips it showed — they ARE the search it ran. Reword more concretely (name the field, methods, populations), or fall back to the manual filters, which run the identical scoring. And report odd readings with your exact wording — the office can improve the translation.

Loosen one facet or swap keywords — scoring needs at least one hit. Also remember the honest limit: the matcher searches what members wrote. A thin pool on some topic means few members have filled collaboration profiles there yet — be the first, and the next searcher finds you.

Your society's signed-in members, through this feature — same audience as the directory, plus your Need/Offer texts. Never money, never demographics, never anything you didn't type into the collaboration fields.

Because it becomes email to many colleagues — the human gate is what keeps the channel signal, not spam. Approval is usually quick; a rejection comes with a note and you can revise and resubmit.

Depends on your target facets — the office sees the live count before sending and can tell you. No facets targets everyone with an active collaboration profile; adding facets narrows it to genuine fits.

The investigator sees you (with your note and contact) on their My-calls view and gets a courtesy email. Next contact comes from them, person to person.

Mark it Filled or Withdrawn and post afresh — posts are short enough that re-posting beats edit history. They expire at 90 days regardless, keeping the board current.

It's an optional add-on — your society may not offer it (yet). If your colleagues at another society mention it, that's why; ask your office about plans.

It queries ClinicalTrials.gov live, but shows the closest dozen matches per scan, not everything — the page's job is 'which studies matter to me,' not a registry mirror. The card links and the keep-digging rail take you to the full registries whenever you want the ocean instead of the catch.

Those come from the international registry cache the office maintains from WHO ICTRP exports — the world beyond ClinicalTrials.gov. Their links open the study's home registry (ISRCTN, ANZCTR, and so on). If you never see such cards, the cache may simply be empty or your topic US-centric; international freshness follows the office's last import.

It's Ralph's reading of registry data that encodes intent inconsistently — deliberately labeled as an AI reading, not a fact. Use it to triage; verify with the study's contact before acting. 'May be adding sites' means exactly that: may.

They're composed from your research profile plus your IntelliConnect collaboration profile. Sharpen either (especially diseases/populations on the collaboration profile) and revisit — themes recompose when your profile changes. The ask box is also a fine override: just say what you want.

Working as designed if nothing genuinely new was posted for it — quiet months send nothing. Check the scan's monthly-email toggle is ON in your saved list, and remember the first email can take up to a month from saving. 'Run now' shows current results any time.

The digest reports newly posted studies; the email itself carries the links. On the page, run the saved scan — the study will rank among results if it still matches your filters. The email is the alert; the page is the explorer.

Live results keep studies with at least one site outside the US and Canada; cached international results filter by their countries. It's the 'what's happening abroad' lens.

Nothing yet — it opens a prefilled study-call form you complete and submit. The call still goes through office review before any colleague is emailed, and responses come to you. It's a head start, not a trigger.

Because the scan reads it: your diseases/populations and datasets make the themes concrete (“CDH infants” beats “neonatology”). Ten minutes on the IntelliConnect profile sharpens every future scan — and makes you findable to collaborators as a bonus.

Correct and deliberate: those sites don't accept prefilled searches, so the rail opens their search page and gives you the phrase on a Copy button — paste and go. PubMed does accept one, so it opens already searched.

It's an optional add-on — your society may not offer it (yet). Ask your office about plans if it would matter to your work.