“Archiving, deduplication, and reports handled by AI. We're back to doing deals.”

The context

LIFTT is a Venture Capital holding company focused on deep tech, with an operational, multi-vertical approach geared toward creating impact.

LIFTT’s edge lies in the quality of its deal pipeline: seeing the right deals first, deciding fast, and maintaining consistent relationships with founders.

The real problem

Deal flow had outgrown the processes managing it. The team was living across:

• CRM as the central system, riddled with duplicate records from multiple ingestion sources
• Outlook email as the primary channel with founders and deal teams, but with no structured archiving
• OneDrive files for pitch decks, attachments, and term sheets — with no automatic link to CRM records
• A monthly report compiled by hand at the start of each month (hours of human work)

The fallout: time-consuming manual deduplication (has this person been contacted? Have they not?), emails that couldn’t be found when due diligence required them (compliance issue), and a monthly report that everyone dreaded.

What the AI agent did

Soraia built an AI agent that autonomously runs the backbone of the deal flow:

1. Monitors incoming emails across deal team mailboxes
2. Identifies deal-related contexts (vs. spam, newsletters, internal communications)
3. Automatically archives eml files and attachments to OneDrive with structured metadata (deal name, founder, date, attachment type)
4. Merges duplicate deals in the CRM: recognises that “Mario Rossi - Startup A” and “M. Rossi (CEO StartupA)” are the same record, and proposes the merge to the team
5. Generates the monthly report with auto-updated charts: deal pipeline, conversion rate by stage, activity by deal team, alerts on stagnant deals
6. Sends alerts on deals that need attention (no contact in X days, term sheet pending, etc.)

The human team only steps in for decisions that require judgment: approving a merge, deciding whether to advance a deal to the next stage, writing a personal message to a founder.

The results

100% of emails archived with attachments and metadata on OneDrive, ready for any due diligence.

Zero hours of manual deduplication: the agent detects and merges duplicates; humans approve the merge in one click.

Monthly report auto-generated at the start of each month, complete with charts and operational recommendations.