AI HR assessment platform
How do I turn 13 years of spreadsheets, Word docs, and scattered email threads into a single AI-powered HR assessment platform — in just 16 weeks? Read the case study to understand how the platform was restructured into scalable, AI-assisted workflows through product thinking, UX systems, and operational design.
Role- Lead UI/UX Designer (Freelance)
Breakfree Consulting Mumbai, India.
Project Timeline Nov 24- Sept 25
Designing an AI-powered HR assessment platform for India's leading consulting firm
Breakfree Consulting ran 13 years of enterprise assessments entirely on Word docs and email threads. I was brought in to change that — and had 16 weeks to ship an MVP that 500+ enterprise clients could actually use.
85%
Reduction in report turnaround time
100%
Audit trail coverage across all actions
4→1
Platform with 3x Faster assessment creation
Why?
I designed AI that suggests, never decides.
Trainers were skeptical of AI-generated content, worried it would feel like losing authorship of their programs. I built a review-before-publish loop where AI drafts and trainers edit, then the system scores quality without removing human judgment.
Result: Adoption was immediate. Trainers felt like collaborators, not users being replaced.
The core Problem — Assessors juggled Word documents, personal spreadsheets, and long email chains to collect scores, compile reports, and communicate findings. There was no single source of truth. No standard framework. Each assessment program was essentially rebuilt from scratch draining time, introducing errors, and making it impossible to scale.
How things were going Earlier?
Different tools used every cycle- Assessors switched between Word, spreadsheets, and email with no consistent workflow—making every cycle a relearning exercise.
Days to compile one report-
A process that should have taken hours was consuming an entire workweek or more than that on every single engagement, without exception.
Standardised frameworks-
Every client operated on a different competency scale, different descriptors, and different scoring logic no tool could hold all of them.
WHY I Understood the domain first before designing the solution?
The goal was clear
Replace a fragmented, manual assessment process with one platform flexible enough to fit any company's way of working. But I didn't start with a brief. I started with Andrea — the person who had run all of this for 20 years.
How I understood each process? — I treated early sessions with stakeholders less like user interviews and more like apprenticeship. Andrea shared internal documents, past engagement reports, and the operational nuances behind each client's framework. Those sessions gave me a vocabulary that made every subsequent design decision land more accurately with the team.
Why?
I designed AI that suggests, never decides.
Trainers were skeptical of AI-generated content, worried it would feel like losing authorship of their programs. I built a review-before-publish loop where AI drafts and trainers edit, then the system scores quality without removing human judgment.
Result: Adoption was immediate. Trainers felt like collaborators, not users being replaced.
The core Problem — Assessors juggled Word documents, personal spreadsheets, and long email chains to collect scores, compile reports, and communicate findings. There was no single source of truth. No standard framework. Each assessment program was essentially rebuilt from scratch draining time, introducing errors, and making it impossible to scale.
85%
85%
Reduction in report turnaround time
Reduction in report turnaround time
100%
100%
Audit trail coverage across all actions
Audit trail coverage across all actions
4→1
4→1
Platform with 3x Faster assessment creation
Platform with 3x Faster assessment creation
How things were going Earlier?
Different tools used every cycle- Assessors switched between Word, spreadsheets, and email with no consistent workflow—making every cycle a relearning exercise.
Days to compile one report-
A process that should have taken hours was consuming an entire workweek or more than that on every single engagement, without exception.
Standardised frameworks-
Every client operated on a different competency scale, different descriptors, and different scoring logic no tool could hold all of them.
Why?
I designed AI that suggests, never decides.
Trainers were skeptical of AI-generated content, worried it would feel like losing authorship of their programs. I built a review-before-publish loop where AI drafts and trainers edit, then the system scores quality without removing human judgment.
Result: Adoption was immediate. Trainers felt like collaborators, not users being replaced.
The core Problem — Assessors juggled Word documents, personal spreadsheets, and long email chains to collect scores, compile reports, and communicate findings. There was no single source of truth. No standard framework. Each assessment program was essentially rebuilt from scratch draining time, introducing errors, and making it impossible to scale.
How things were going Earlier?
Different tools used every cycle- Assessors switched between Word, spreadsheets, and email with no consistent workflow—making every cycle a relearning exercise.
Days to compile one report-
A process that should have taken hours was consuming an entire workweek or more than that on every single engagement, without exception.
Standardised frameworks-
Every client operated on a different competency scale, different descriptors, and different scoring logic no tool could hold all of them.
WHY I Understood the domain first before designing the solution?
The goal was clear
Replace a fragmented, manual assessment process with one platform flexible enough to fit any company's way of working. But I didn't start with a brief. I started with Andrea — the person who had run all of this for 20 years.
How I understood each process? — I treated early sessions with stakeholders less like user interviews and more like apprenticeship. Andrea shared internal documents, past engagement reports, and the operational nuances behind each client's framework. Those sessions gave me a vocabulary that made every subsequent design decision land more accurately with the team.
WHY I Understood the domain first before designing the solution?
The goal was clear
Replace a fragmented, manual assessment process with one platform flexible enough to fit any company's way of working. But I didn't start with a brief. I started with Andrea — the person who had run all of this for 20 years.
How I understood each process? — I treated early sessions with stakeholders less like user interviews and more like apprenticeship. Andrea shared internal documents, past engagement reports, and the operational nuances behind each client's framework. Those sessions gave me a vocabulary that made every subsequent design decision land more accurately with the team.
Working from a formal SRS as a design constraint, not just a reference
The Software Requirements Specification forced precision in every interaction data model and pattern. Where other designers might treat it as background documentation, I treated it as my primary design constraint — ambiguity in requirements translated directly to ambiguity in the UI. This discipline saved multiple costly rounds of revision downstream.
And now here comes the moment of making DECISION ON SCREEN where all these things made sense
Working from a formal SRS as a design constraint, not just a reference
The Software Requirements Specification forced precision in every interaction data model and pattern. Where other designers might treat it as background documentation, I treated it as my primary design constraint — ambiguity in requirements translated directly to ambiguity in the UI. This discipline saved multiple costly rounds of revision downstream.
And now here comes the moment of making DECISION ON SCREEN where all these things made sense
Competency mapping — turning subjective judgment into structured, scorable frameworks
Competency mapping — turning subjective judgment into structured, scorable frameworks
Breakfree's core differentiator is their competency mapping depth. The platform needed to digitise this without flattening the nuance supporting hierarchical competencies, sub-competencies, custom scoring scales, and persistent library management.
I designed a three-layer architecture:
Competency Library (reusable across engagements) → Competency Maps (assembled per designation) → Assessment Scoring (applied per candidate). Each layer has its own CRUD workflow, but they interconnect in a way that makes reuse the default path — not a workaround.
The platform replaced a process that took weeks with one that takes minutes.
Ux Decision
Progressive Disclosure in Assessment Creation
Instead of exposing all assessment options upfront, the wizard pattern guides trainers through name → type → components → schedule, reducing cognitive load at each step and preventing incomplete submissions
Compliance
Audit-Ready Data Architecture
Every user action — from login to score modification is timestamped and attributed. The RBAC model ensures data access follows the principle of least privilege, critical for GDPR compliance and enterprise security audits.
AI Ethics
Human-in-the-Loop AI Content
AI-generated content never auto-publishes. Trainers must explicitly review, edit, and approve every piece. The evaluation scoring provides transparency into content quality without removing human judgment from the loop.
Scalability
Modular Content Management System
Assessments, report structures, AI profiles, and competency maps are independent modules that interconnect. This modularity lets Breakfree serve diverse industries without rebuilding the core platform for each new client vertical.
The platform replaced a process that took weeks with one that takes minutes.
The platform replaced a process that took weeks with one that takes minutes.
Ux Decision
Progressive Disclosure in Assessment Creation
Instead of exposing all assessment options upfront, the wizard pattern guides trainers through name → type → components → schedule, reducing cognitive load at each step and preventing incomplete submissions
Compliance
Audit-Ready Data Architecture
Every user action — from login to score modification is timestamped and attributed. The RBAC model ensures data access follows the principle of least privilege, critical for GDPR compliance and enterprise security audits.
AI Ethics
Human-in-the-Loop AI Content
AI-generated content never auto-publishes. Trainers must explicitly review, edit, and approve every piece. The evaluation scoring provides transparency into content quality without removing human judgment from the loop.
Scalability
Modular Content Management System
Assessments, report structures, AI profiles, and competency maps are independent modules that interconnect. This modularity lets Breakfree serve diverse industries without rebuilding the core platform for each new client vertical.
Ux Decision
Progressive Disclosure in Assessment Creation
Instead of exposing all assessment options upfront, the wizard pattern guides trainers through name → type → components → schedule, reducing cognitive load at each step and preventing incomplete submissions
Compliance
Audit-Ready Data Architecture
Every user action — from login to score modification is timestamped and attributed. The RBAC model ensures data access follows the principle of least privilege, critical for GDPR compliance and enterprise security audits.
AI Ethics
Human-in-the-Loop AI Content
AI-generated content never auto-publishes. Trainers must explicitly review, edit, and approve every piece. The evaluation scoring provides transparency into content quality without removing human judgment from the loop.
Scalability
Modular Content Management System
Assessments, report structures, AI profiles, and competency maps are independent modules that interconnect. This modularity lets Breakfree serve diverse industries without rebuilding the core platform for each new client vertical.
more projects
Ready to Make Your Brand success and Unforgettable?
Let's make it happen!
I'm open to full-time product design roles where AI, compliance, and complex systems are the actual brief — not an afterthought. If your product has to work in the real world, under real pressure, for real users — that's exactly where I do my best work.
Open to roles
Sakshi Gupta
Fulltime
Skills and Tools
Web Design
Mobile Applications
Saas
Design Systems
AI workflow UX
UX Research
Visual Design
What I bring
Compliance-first thinking
AI product experience
Complex systems
Cross-functional collab
Ready to Make Your Brand success and Unforgettable?
Let's make it happen!
I'm open to full-time product design roles where AI, compliance, and complex systems are the actual brief — not an afterthought. If your product has to work in the real world, under real pressure, for real users — that's exactly where I do my best work.
Open to roles
Sakshi Gupta
Fulltime
Skills and Tools
Web Design
Mobile Applications
Saas
Design Systems
AI workflow UX
UX Research
Visual Design
What I bring
Compliance-first thinking
AI product experience
Complex systems
Cross-functional collab
Ready to Make Your Brand success and Unforgettable?
Let's make it happen!
I'm open to full-time product design roles where AI, compliance, and complex systems are the actual brief — not an afterthought. If your product has to work in the real world, under real pressure, for real users — that's exactly where I do my best work.
Open to roles
Sakshi Gupta
Fulltime
Skills and Tools
Web Design
Mobile Applications
Saas
Design Systems
AI workflow UX
UX Research
Visual Design
What I bring
Compliance-first thinking
AI product experience
Complex systems
Cross-functional collab