SpeakUP is the ethical AI interview-training platform for universities — multimodal feedback on speech, voice and presence, designed for the way career services actually work.
Every year, hundreds of thousands of finalists graduate into a recruitment process they have never rehearsed against. The career office is small. The cohort is not.
Students rehearse in front of a mirror or skip rehearsal entirely. The first real interview becomes the practice run — and the cost of failure is the offer.
Friends say "you did great." That's not a baseline. Without signal on speech, pacing, body language and content, students improve by chance, not by craft.
A two-person team cannot run mock interviews for two thousand finalists. So they don't — and the students who need help most are the ones who never book a slot.
SpeakUP turns every screen into a private interview booth. No appointment, no judgment, no waiting list — and a report waiting on the other side.
Behavioral, technical, case, language switch. Pick role, sector and difficulty — or load your faculty's question bank.
Real questions, real time pressure, real camera. No actors, no scripts, no second takes — just like the day that counts.
Verbal, vocal and non-verbal scores within seconds. Concrete moments to revisit. A measurable target before the next attempt.
One dashboard per school. Adoption you can audit. Outcomes you can put in front of a rector.
students in Portuguese higher education. Your scale problem, already solved.
Cohort-level adoption, performance and progress, faculty by faculty.
Verbal, vocal and non-verbal — auditable models, bias-tested quarterly.
SAML, OAuth, your IdP. Students sign in once with their university account.
Rubrics co-designed with senior recruiters at a global leader in talent.
Not one shot at a recruiter's calendar. As many takes as it takes — at 2pm or 2am, in the bedroom, in pyjamas, without anyone watching.
"I did fourteen mock interviews before my real one. The first three were rough. By the fourteenth, I knew exactly what to fix."
"The non-verbal feedback was the part I didn't know I needed. I had no idea I was looking down every time I talked about my project."
"My career office had two slots a month. I had thirty interviews to prepare for. SpeakUP filled the gap nobody else could."
Other tools are built for HR teams and re-skinned for schools. SpeakUP starts at the university and earns its trust there.
Roles, dashboards, procurement cycles and pricing models designed for higher education — not retrofitted from a corporate product.
Three modalities, transparent rubrics, regular bias audits. Students see what the model evaluated, and why.
EU hosting, automatic recording deletion, student-controlled sharing. Nothing leaves the institution without consent.
Co-built with senior recruiters at Hays. The rubric reflects what real interviewers actually weigh in real interviews.
Annual licenses, per institution, billed once. No per-student creep. No surprise overage when a cohort grows.
Short, honest, on the record. If your legal team has a sixth question, we have an answer for that one too.
Every model we deploy is audited quarterly against gender, accent and age slices, with results published to subscribing institutions. We do not score appearance, attractiveness or perceived demographics — and we never will. Audit reports are available under NDA before pilot.
SpeakUP is hosted in the EU, processes data under a signed DPA, and gives every student granular control over their recordings — including a one-tap permanent delete. We are an ISO 27001-aligned processor; the institution remains the controller.
14 weeks, from kickoff to debrief. Three of those are setup, ten are live with students, one is for evaluation. We deliver an outcomes report at week 13 with adoption, performance lift and qualitative student feedback.
Yes. Native integration with Moodle, Canvas and Blackboard via LTI 1.3. SSO via SAML 2.0 and OAuth 2.0. Roster sync is supported via SCIM or scheduled CSV — your IT team picks the one they already use.
By default, recordings are deleted automatically 30 days after analysis. Scores and aggregate trend data are retained for the duration of the contract. Students can opt-in to longer retention for portfolio purposes; nothing is sold, shared or used to train third-party models.
SpeakUP is classified, documented and operated as a limited-risk educational system under the EU AI Act. We provide model cards, risk assessments and a human-in-the-loop pathway for any scoring a student wishes to contest. We do not deploy in employment-decision contexts.
Pilot in one faculty. Measure the lift. Roll out the rest when the data is yours.
