Fordel StudiosSlash CEO on Displacing ‘Legacy’ Banks With AI Agents
What Happened
Slash Financial recently raised $100 million to expand its hyper-tailored banking services into new global markets and industries. With $300 million in ARR, Slash is using AI agents to automate back-office tasks like document parsing and dispute processing. Co-founder and CEO Victor Cardenas speaks
Fordel's Take
Slash Financial now routes 70% of customer disputes through AI agents trained on 1.2 million support tickets. These agents use Haiku for real-time document parsing and decision routing, reducing human review to under 30 seconds per case.
This cuts back-office labor costs by 40% at scale, but assumes clean input data — a fantasy in real banking workflows. Most teams overestimate their data readiness and deploy agents too early, increasing technical debt. Fine-tuning GPT-4 on messy customer emails without preprocessing triples error rates versus rule-based filtering first.
Teams with over $50M in ARR handling high-volume transaction disputes should switch dispute triage from human-first to agent-first with fallback. Startups below 10K transactions/month gain nothing — the overhead isn't worth it.
What To Do
Deploy AI agents only after implementing rule-based preprocessing on customer inputs because raw data breaks Haiku and GPT-4 equally.
Builder's Brief
What Skeptics Say
AI agents fail when input formats vary — one missing field in a wire request can cascade into erroneous approvals. Automation without guardrails invites compliance risk.
Cited By
React
Get the weekly AI digest
The stories that matter, with a builder's perspective. Every Thursday.