New program
Build a program by answering a few questions
In New program it asks one question at a time: who takes part, how long, who chooses the matches. Then the program’s own assistant finishes the setup.
Your mentoring program, on autopilot
It watches the program, does the work and keeps the term on its calendar. You choose what it may do on its own; everything else waits for your approval.
You set the level. Anything above it waits on a card for you.
Autonomy level
Deleting and removing members never run by rule, at any level.
Give it the term’s dates. It ties each step to its date, has the work ready on time and waits for you where you want it to.
Applications open
Invitations to 140 people from your spreadsheet
Applications close
The form closes itself at the end of the day
Matching
42 pairs on one card, each with its score
Kickoff
The kickoff event and first to-do list, drafted
Midterm
A message in the app to 6 quiet matches
Closing event
Certificates for 38 people, the end check-in open
New term
54 members invited back, finished mentees as mentors
Each of these is in the admin panel now, in matched programs and in pool programs where mentees book mentors themselves.
New program
In New program it asks one question at a time: who takes part, how long, who chooses the matches. Then the program’s own assistant finishes the setup.
Cohort matching
Ask it to match the new cohort. It scores every pair with your matching rules and brings up to 50 pairs on one card, each with the rules it meets.
Program calendar
Tell it the term’s dates: applications, kickoff, midterm, the closing event. It plans dated tasks around them and closes the application form on its deadline.
Program inbox
Messages sent to the program, and answers to its own e-mails, come to the assistant. It drafts the reply and sends it as a message in the app, from the program.
From a spreadsheet
Attach the list in the chat. It maps the columns to e-mail, name and your profile questions, and adds up to 100 people per card, with or without an invitation.
Notifications
When a scheduled task leaves cards for you, cannot run, or credits run low, the right administrator gets an e-mail. You do not have to keep the chat open.
New term
It copies the program for the new term without its members, then invites the ones coming back: mentors, mentees who continue, and mentees who finished, as mentors.
Matching
Ask for the best mentors for a mentee or a team. It scores them with your matching rules and names the rules each one meets and misses.
Pool programs
Where nobody is matched and mentees book mentors themselves, it picks two or three mentors per venture by profile, free slots, earlier meetings and your tag rules.
Mentor time
It shows mentors who never added hours, ran out, or are fully booked, and drafts the right ask for each group. Paste a mentor’s reply and it turns it into hours on their calendar.
Participants’ requests
When participants’ support requests wait for the program, a box above the chat lists the oldest ones. Solve one, or hand them all to the assistant at once.
Several programs
In the chat you pick the programs it works on, grouped by customer. A shared announcement is posted in each one and mailed once: someone in four programs gets one e-mail.
Reports
Ask about the program in your own words and it answers from the live data with figures and charts. On any report, "Explain with AI" gives you its findings.
Presentations
One presentation per program type: community, employee, startup, student and alumni. Open it, present it, share it.
Mentornity AI for Alumni Associations
Mentornity AI for Communities and Associations
Mentornity AI for Employee Mentoring
Mentornity AI for Nonprofits and Foundations
Mentornity AI for Professional Associations and Business Networks
Mentornity AI for Startup Programs
Mentornity AI for Student and Alumni Mentoring
Mentornity AI for Women’s Mentoring ProgramsAutonomy level
The account owner sets one level for all programs. Scheduled tasks work within it; anything above it comes to you as a card.
Every step comes as a card. Nothing is applied until an administrator approves it.
Tasks may do limited things on their own:
Everything in safe rules, plus:
Deleting and removing members never run by rule, at any level. Every rule has a limit per run, and each change is checked again just before it is applied.
Kinds of mentoring
It carries a playbook for seven kinds of mentoring: the profile questions to ask, the posts, forms and surveys worth proposing, what to measure and the usual pitfalls.
Suggestions only, safe rules or full autopilot. Anything outside the level waits on a card.
The model sees participants under aliases, never their contact details.
Use credits, or your own Anthropic API key and Anthropic bills you.
Participants can connect Mentornity to Claude, ChatGPT and other AI assistants over MCP, to find mentors, plan meetings and fill in forms. It only reads by default and shows a preview before any change.
Mentornity AI is the assistant in Mentornity’s admin panel. It watches a mentoring program, does the work (matches, meeting suggestions, e-mails, rules, mentor hours, reports and certificates) and brings each change as a card the administrator approves.
Only what you allow. On suggestions only, every change waits on a card until an administrator approves it. On safe rules or full autopilot, scheduled tasks apply the kinds of changes the level covers, within limits per run; everything else still waits on a card.
Yes. There it never proposes matches; it suggests which mentors each mentee or venture should meet, by profile, free slots, earlier meetings and tag rules, keeps mentor hours open and sets the tag rules that decide who can book whom.
The model sees participants under aliases and never their contact details. The real names are put back in by Mentornity when you read the answer.
With credits, or with your own Anthropic API key, in which case Anthropic bills you and no credit is used. A "with AI" button spends nothing until you send its request.
Yes, through the Mentornity connector: they connect Mentornity to Claude, ChatGPT or another assistant over MCP and use it as themselves, with a preview before every change.
Yes. In the chat you pick the programs it works on, from every program in your account. A shared announcement is posted in each of them and mailed once, so someone in several programs gets one e-mail with every text in it.
The mails it proposes carry its own reply address. The answer comes back to the assistant, shows in the participants’ requests box and you get an e-mail. Hours a mentor gives become a card you approve.
Within the level you choose. On safe rules, scheduled tasks accept applications, send follow-up e-mails, book meetings in open slots and answer support requests. On full autopilot they also match pairs scored 70% or more, end inactive matches, publish certificates and send messages in the app. Deleting or removing members always waits for an administrator.
You give it the program calendar: the start and end of the term, when applications open and close, the kickoff, the midterm and the closing event. It plans dated tasks around them and can close the application form on its deadline.
By e-mail: when a scheduled task leaves cards for you, when a task could not run and why, and when credits run low.
Bu sayfa Türkçe olarak da mevcut.
Türkçe devam etCette page est aussi disponible en français.
Continuer en françaisDiese Seite gibt es auch auf Deutsch.
Auf Deutsch fortfahrenEsta página también está disponible en español.
Continuar en españolQuesta pagina è disponibile anche in italiano.
Continua in italianoEsta página também está disponível em português.
Continuar em portuguêsDeze pagina is ook beschikbaar in het Nederlands.
Doorgaan in het Nederlands