Featured Case Study
Revolutionizing École des Ponts Business School Admissions with AI
École des Ponts Business School Case Study
Revolutionizing a leading business school's admission process with AI
The Challenge
Every year, École des Ponts Business School, one of Europe's top ranked business schools, processes thousands of applications, CVs, identity documents, transcripts and test reports. Until now, this was fully a manual handling process resulting in processing times often stretching to several weeks or more. This resulted in escalating operational costs through repeated human validations and rechecks and exposed the school to human error and data inconsistencies.
The Solution
In March 2025, École des Ponts Business School launched a pilot with Semantikmatch to automate the parsing and analysis of application dossiers. Our AI platform was deployed successfully and during the pilot phase successfully extracted candidate identity data with 100% accuracy, identified professional experience with 96% precision, and assessed academic criteria according to admissions guidelines at a reliability rate between 96% and 98%.
Our Impact
On a sample of 500 applications, these performance levels have radically accelerated the review cycle: what once required up to 20 minutes of manual evaluation is now completed in mere seconds, all while delivering 10× the accuracy of manual data entry and eliminating errors, bias and omissions.
Optimized Candidate Experience
Beyond internal efficiency gains, Semantikmatch elevates the applicant journey. Candidates gain access to a real-time portal showing the completeness of their file, with instant notifications for any missing documents or clarifications needed.
Next Steps
Thanks to the success of the pilot project, École des Ponts Business School plans to roll out the solution across all its programs (MSc, MBA, EMBA) with full GDPR Article 22 compliance.
"We are convinced that well-configured and well-governed AI will become a true co-pilot for our admissions juries, ready to take selection to new heights of speed, fairness and confidence"— Product Manager, École des Ponts Business School
FAQs
Frequently Asked Questions
How does Semantikmatch handle international transcripts and grading systems?
We support 70+ national grading systems (and their variants). We distinguish institution-specific scales (e.g., Harvard vs. UCLA vs. Concordia). Grades are normalized into comparable indexes you define. Ambiguous or edge cases are flagged for manual review. Benefit: 90% reduction in processing time with 96–100% accuracy on benchmark samples.
Do you perform behavioral analysis on video interviews?
No. We do not decode emotions or expressions (prohibited under the AI Act). Instead we: Transcribe the spoken content, apply a custom rubric you define, flag potentially scripted or AI-generated answers, and assess English proficiency using phonetic and syntactic indicators (~90% accuracy). 70–80% of clearly strong or weak responses can be processed without human review.
How is Semantikmatch different from internal scoring sheets (Excel, Google Sheets)?
We read and verify each document directly (transcripts, diplomas, letters). We detect inconsistencies between declared data and attached documents. We automate verified scoring and flag anomalies. Benefit: 80–90% reduction in manual verification time without changing your business logic.
How does Semantikmatch help with legal risk and subject access requests (SARs)?
Full GDPR compliance, including Article 22 rights. Human review is always possible. Traceable evaluations at the document level with exportable audit logs. You get all the proof you need to handle SARs quickly and confidently.
What is the implementation timeline and integration process?
Rollout in 4–6 weeks for a first program: scoping workshop, Slate integration, user onboarding. Plug-and-play integration with Slate and other ATS: scores appear directly in your existing interface.
Can I connect Semantikmatch to my CRM/ATS (Slate, Salesforce, HubSpot, Workday, Taleo, etc.)?
Yes: dedicated connectors for Slate, Greenhouse, Workday, Taleo, etc. Two-way sync: applications flow in; decisions flow out in real time, no manual entry.
What happens with borderline applications (edge cases)?
You set your thresholds; we then: Highlight GPAs out of range, missing data, or conflicting information. Group these cases in a dashboard for manual review.
How do exports and reporting work?
All scores, statuses, and comments are pushed to Slate or your ATS. Audit-ready reports per candidate or cohort, exportable as PDF or Excel.
How does Semantikmatch differ from CRM-native AI features (e.g., Slate AI)?
Slate AI: assists individual document review and relies on human interaction at each step. SemantikMatch: a full agent system that extracts and scores admission criteria across all candidates automatically, applies your eligibility thresholds without user input, logs every decision for audit and compliance, surfaces only edge cases to staff, saving 80% of each officer's workload.
Where and how are my data stored and protected?
European hosting: all your data (CVs, reports, logs) are stored in ISO 27001-certified data centers in France and Germany. End-to-end encryption: data encrypted in transit (TLS 1.2+) and at rest (AES-256). Restricted access: server access controlled via IAM and multi-factor authentication. Regular backups: daily archives with GDPR-compliant retention. Security audits: quarterly third-party audits and penetration tests ensure data integrity and confidentiality.