Case Study: Internal Operations
Let's look at why and how we implemented a Credit Analyst use case for a London based fintech Scale Up
NEW AUTOMATION OPTIONS
AI enables computers to handle unstructured data, unlocking the automation of complex new processes
Deterministic workflow
Agentic workflow
Agent
Agent freely deciding which tools to use based on his goal
BEST PRACTICES
Implementing agentic workflows requires intelligently integrating AI into existing human-driven processes
Find the parts of the overall process which fit well for AI
Establish “Human-in-the-loop” approach with proper escalation
Prioritize use cases based on business case, AI output quality & implementation speed
Plan for iterations: the hard part is to get the evaluations right (=output quality & reliability)
Establish tech stack, tools & guardrails so people can experiment quickly securely & compliantly
CREDIT ANALYST EXAMPLE
Deep dive: The agentic workflow for our credit analyst
Outcome
SAVINGS / LEAD
5 EUR
(20 min saved)
Cost to run
0,20 EUR
Net Savings / lead
4,80 EUR
Monthly Savings
1.920 EUR
(assuming 400 leads / month)
Plus increased quality since the AI acts as a second pair of eyes
PROJECT APPROACH
We built quickly with n8n, evaluated it with a eval data set, tested in the real world via a web app and then integrated into their CRM
DEMO
Try it out: Test the agentic workflow live
Unlock full results
Enter your email to see the complete analysis
By entering your email, you agree to our Privacy Policy