Cutting Manual Email Triage by Over 70%: An AI Transformation for Jome
How Jome built an AI-powered intake and response system with Claude — automating client email processing, personalized drafting, and finance reporting in an AI Transformation engagement with Easyflow.
Industry
Services
AI Transformation
Timeline
Phased Engagement
Team Size
6 specialists
Jome helps homebuyers search, compare, and purchase new construction homes in one place, bringing together pricing, incentives, and builder details. As client volume grew, two problems surfaced: response quality suffered, and financial data piled up in spreadsheets. Jome brought in Easyflow to build Claude agents into these workflows. The agents handle email triage, draft responses that match Jome's voice, and pull financial signals directly into decisions. Manual email handling dropped 70%.
70%+
Manual email triage reduced
3
AI agents deployed
100%
On-time delivery
Challenges
Scaling Client Communication
Every incoming client email needed reading, interpreting, and categorizing before a reply could go out. As volume grew, this step became the bottleneck to a fast, high-quality response.
Turning Data into Decisions
Financial and operational data lived in exports and reports. The team wanted to move faster from data to decision, spotting patterns, anomalies, and seasonality without manual review standing in the way.
The Solution
Easyflow's AI Transformation methodology starts with an audit: mapping the operation, ranking opportunities, and scoping the highest-impact automations. For Jome, that audit surfaced three priorities: automating email intake, improving response speed and quality, and turning financial data into insight the team could act on. These became the phases of the engagement.
Audit & Scope
Easyflow mapped Jome's client communication and finance reporting workflows and prioritised the phases of the engagement around email intake, response quality, and financial insight.
Build the Email Parser and Intent Detector
A Claude-based AI agent was built to scan incoming client emails, extract relevant details, and detect intent, so requests are categorized and routed automatically.
Build the Personalised Response Assistant
A second agent was trained to draft context-aware replies based on the detected intent, giving the team personalised responses ready to send.
Connect Detection to Response
The two agents powered by Claude were linked to a detected intent flows directly into a matching draft reply, closing the gap between triage and response.
Build the Finance Operations Insight Agent
An internal Claude AI assistant was built to analyse financial and operational exports, surface patterns, anomalies, and seasonality, and answer team questions in plain language.
Handover & Ongoing Iteration
Easyflow stayed responsive after delivery, adjusting the agents quickly as Jome's processes and data evolved.
What We Built
01
Email Parser and Intent Detector
Scans incoming client emails, extracts relevant information, and classifies intent so requests route automatically.
02
Personalised Response Assistant
Drafts context-aware replies based on detected intent, giving the team personalised responses ready to send.
03
Intent-to-Reply Automation Pipeline
Links detection and drafting into a single flow, taking a client email from intent to a ready reply with minimal manual review.
04
Finance Operations Insight Agent
Analyses financial and operational exports with Claude LLM to surface patterns, anomalies, and seasonality.
05
Natural Language Query Interface
Allows the finance team to ask questions about the data directly and get plain-language answers.
06
Executive Insight Summaries
Generates concise summaries of key drivers and trends for leadership review, translating data into plain language.
Broader Impact
Client Ops Efficiency
The team spends far less time reading and categorising emails by hand, freeing time for higher-value client interactions.
Faster, Tailored Response
Detected intent flows straight into a drafted reply, helping the team respond quickly with personalised recommendations.
Data-Driven Finance Decisions
Financial patterns and anomalies now surface through natural-language queries and summaries powered by Anthropic's most efficient models.
A Foundation for Further Transformation
The audit-first approach gives Jome a scoped roadmap for extending AI further into client operations and finance.