All projects
Financial Services
Fintech revenue operations
Brief Organic leads were falling while customer acquisition costs rose
Built A revenue operations layer that prioritises leads and customer profiles, guiding outreach and targeting
Result Productivity per FTE rose threefold. The go-to-market team moved from 8 to 3 FTEs, with USD 500,000 in FTE savings
Lender credit modelling
Brief Credit modelling knowledge was concentrated in individuals and difficult to retain across the business
Built Initial credit models using alternative data, with AI inputs supporting iterative improvements
Result The bad debt ratio fell by 5 percentage points
Lender customer growth
Brief High organic acquisition costs were putting pressure on marketing returns
Built Customer lifecycle mapping, WhatsApp conversion and reactivation workflows, paid-channel retargeting, and SEO and AI search optimisation
Result Marketing costs fell by 30% while revenue held constant
Lender reconciliation
Brief Manual month-end reconciliation took around ten person-days
Built Automated reconciliation connected to QuickBooks, supported by AI workflows
Result Accounting and reconciliation effort fell from 10 to 1 person-day
Venture capital fund operations
Brief Fragmented venture partner workflows and manual handovers made it hard to support more than 100 portfolio companies and prepare LP reports
Built Connected workflows and data collection, with AI note-taking and tools to capture information in the right systems
Result 40% of team capacity was released from fund operations and portfolio company support
Venture capital sourcing and diligence
Brief Limited team capacity constrained deal sourcing and the review of incoming pitch decks
Built Automated lead collection and initial deck screening to highlight opportunities aligned with the investment thesis
Result The team screened 60% more companies over the same period
Venture capital LP sourcing
Brief LP connections were fragmented across the team, making it difficult to identify investors for a niche early-stage strategy
Built A shared database of warm LP connections, with investor screening informed by fundraising activity at regional start-ups matching the fund’s thesis
Result Warm LP connections consolidated in one shared database
Accounting workflows
Brief Repetitive accounting work took junior staff time, while existing AI solutions had high token costs and needed further refinement
Built Purpose-built code and a smaller, specially trained model for selected accounting workflows
Result Selected-workflow models achieved 95%+ accuracy, with 80% lower token spend
Legal Services
Barrister case preparation
Brief Understanding documents and conducting legal research required substantial time
Built Automation for data-room summaries, research in Westlaw and LexisNexis, evidence organisation and filing
Result A 30% improvement in efficiency
Xero
QuickBooks
Sage
Salesforce
HubSpot
Attio
Outlook
Gmail
Proton Mail
Excel
Google Sheets
Airtable
SAP
Odoo
Zoho