xEllerator pairs mentees with the right mentors using an equity-aware matching engine, then tracks the goals, sessions and outcomes that close the leadership gap in STEM.
How a match is scoredSources: Women in the Workplace (McKinsey and LeanIn.Org); The Sponsor Effect (Center for Talent Innovation).
Everyone answers a few open questions when they join. Amara wrote that she wants her first lead role and needs help with stakeholder presentations. Ingrid wrote that she coaches public speaking. The engine reads these answers and scores every allowed pairing. This pair scored 84.
Compatibility is 62% of the score. The card splits it into five bars. Goal fit is the other 38%: Ingrid scores 91 because she has done what Amara wants to do. The +6 is the equity uplift. It is capped at 12 points and goes to pairs where the mentee needs extra support and the mentor can give it. If compatibility is too low, the uplift is not applied.
Some rules are checked before scoring. The mentor must be at least one level more senior, and their availability must overlap. Dealbreakers are honoured. No mentor gets more mentees than the capacity she set.
The HR console has an Equity page. It splits every mentee by gender identity, career stage and field, and shows 29 measures side by side with the figure for everyone: who got matched and how long it took, how often pairs meet, what the mentees say about the match, what mentors did beyond advice, and what changed by the end.
A group is shown only if it had five people or more when its figures were worked out. A cell that sits well below everyone gets a red mark and a plain sentence in the "Needs a look" list, and is written to the activity log the first time it appears. Answers about promotions, moves and confidence are only counted when the person ticked the box that allows it.
| Group | People | Matched | Days to match | Accepted their match | Score |
|---|---|---|---|---|---|
| Everyone | 49 | 92% | 6 days | 90% | 81 |
| Woman | 31 | 94% | 5 days | 93% | 83 |
| Man | 13 | 92% | 9 days | 77% | 79 |
| Non-binary | under 5 | Fewer than 5 people, so no numbers are shown | |||
| Prefer to self-describe | under 5 | Fewer than 5 people, so no numbers are shown | |||
A sponsor uses their standing to open doors. The app asks mentors, when they join, whether they will make introductions, put someone forward for a stretch assignment, or speak up in a promotion conversation. Mentors who say yes score higher for mentees who have no senior backing.
During the match, a mentor logs each of those things in a few words. The mentee sees the list on her dashboard. The Equity page counts introductions, stretch opportunities and advocacy per match, by group.
ISE&S is a nonprofit research institute in New Jersey that works on social equity. Its board includes researchers at John Jay College (CUNY), Rutgers, Kean University, Suffolk University and the College of Charleston.
Research from its board members drove the idea behind the kind of mentoring xEllerator supports. Dr. Sarah-Lee Boodoo's 2025 study called for formal mentoring programmes and sponsorship for women.
ises-nj.org
DMS is a consulting firm in Metuchen, New Jersey. It works on AI governance and cybersecurity for businesses, governments and nonprofits.
DMS built xEllerator and runs it. Aamir Ahamad wrote the app. He works at DMS on information security and AI governance.
strategydms.comDr. Deborah Mohammed-Spigner is on the ISE&S founding board and is CEO of DMS. The programme team that runs the pilot is six people from DMS and the ISE&S board.
Run one cohort with us. At the end you get a written report on who was matched, how often pairs met, what participants said, and where the gaps were. Then you decide whether to continue.
The app and its database run on Railway, in its US East region in Virginia. All organisations share one database. Every record carries the organisation it belongs to, and HR's screens read only their own organisation's records. Your HR admin sees names. The programme team, the small team running this pilot, sees names across organisations. It also reads check-ins, answers and requests for help pair by pair. Nobody on it reads these for a pair they are in, or for the pairs in an organisation where they are HR. A group's figures are shown only if it had five people or more when they were worked out. Answers about promotions, moves and confidence are counted only when the person ticked the box that allows it. Everything between your browser and xEllerator is sent over HTTPS. Security and data has the details.
The engine proposes. HR releases each proposal, or withholds it with a logged reason; HR cannot change the ranking. The mentee decides first, then the mentor. Either can decline, and that mentor is not proposed to her again.
Yes. An OpenAI model reads what people wrote when they joined and picks out the skills and goals in it. The app compares these by meaning. The score itself is added up by the app's own rules, with the fixed weights shown above. The model then writes the reason on each proposal. It is given only the facts behind the score. Check-in answers, private notes and requests for support are never sent to the model. AI spend has a daily limit, and when the model is off or the limit is reached, the app uses word matching and written templates instead.
One session a month at least, plus short check-ins in the app. Mentors set how many mentees they will take and how many hours a month they can give, and the engine keeps to that.
You get the outcomes report and a recommendation on next steps. There is no automatic renewal and no long-term obligation. Whether to continue is your call.
For mentees, mentors and HR admins. Use the email your invitation was sent to.
Your account knows your role. You land on your own dashboard.