“350k+ candidates on one platform. AI agents handle the operational work so our team doesn't have to.”

Context

Aegis Group (“Driving business through people”) is an Italian Human Consulting group headquartered in Milan, structured across 7 vertical business lines:

• Aegis FSI, financial services
• Human Capital, HR organization development
• Four Corners, luxury, fashion, beauty, design, lifestyle
• Greentalent, energy & sustainability
• Value Stream, industrial automation & manufacturing
• HyperSet, technology & digital
• Glinters, startup & scaleup

Each line has its own team, its own open roles, and its own rules for working with candidates. Together, they manage a database of over 350,000 candidates.

The real problem

The pre-Soraia situation was fragmented and operationally expensive:

• Data scattered across spreadsheets, with overlaps and duplicates impossible to catch manually.
• A legacy system treated as a burden: “ugh, now I have to fill in the system” — a chore rather than a useful tool for day-to-day recruiting work.
• Fully manual candidate management, from CV intake to client presentation: parsing by hand, qualification by hand, follow-up by hand.
• Zero consolidated analytics: producing even a single basic group-level metric (active candidates, average time-to-fill, conversion rate) required hours of exports and manual cross-referencing between files.

The result: senior recruiters and managers were burning high-value time on data entry and reconciliation instead of advising candidates and clients.

What we built

Soraia built a custom recruitment platform with AI agents embedded in the operational workflow — replacing spreadsheets and the legacy system as the single source of truth for all 7 group lines:

1. AI CV parsing agent — receives a CV (PDF, Word, image), extracts structured data with far greater accuracy than manual parsing: skills, experience, languages, certifications, previous roles — all in 30 seconds.
2. Cross-brand deduplication — the agent recognises when the same profile comes in from different lines and suggests a merge, preserving the full interaction history with each division.
3. Multi-brand workflow with granular permissions — each manager sees and manages candidates within their own line, with cross-divisional visibility where needed.
4. Consolidated analytics — dashboard with group-wide metrics: active candidates, average time-to-fill by role, conversion rate by line, sourcing channel breakdown. No more Excel exports.

Automatic candidate↔role cross-line matching is on the roadmap: recruiters already benefit from parsing, deduplication, and unified visibility, but the system doesn’t yet proactively suggest matches. That’s the next step in AI integration.

Results

350,000+ candidates on a single platform, all searchable through the same interface, with consistent history, notes, and pipeline status across lines.

7 lines aligned on the same system, each with its own specific rules running on shared infrastructure.

The system is no longer a burden: AI agents absorb the operational work (parsing, deduplication, reconciliation), giving recruiters a tool that actually works for them.

Real-time analytics — no more hours of exports and data reconciliation just to produce a single number.

Where we’re headed

The platform is live and we continue to iterate with regular sprints to embed more AI into the group’s specific workflows and free up more recruiter time. Next areas: automatic candidate↔role matching, agent-assisted candidate outreach, and dynamic pipeline scoring.