Enterprise Knowledge Search · IT Operations Automation
AppFolio needed a natural-language interface that could help property managers query portfolio state, execute bulk actions, and troubleshoot complex workflows while preserving low latency, observability, and prompt quality as the assistant grew more capable.
Reported impact
Realm-X users save meaningful time on day-to-day property-management work, AppFolio improved text-to-data accuracy, and the team gained a monitored agent workflow that can scale new actions and data models without losing visibility.
AppFolio built Realm-X Assistant as an embedded generative-AI copilot for property managers and started on LangChain so model providers, tool calling, and structured outputs could be swapped without rewriting the app. As requests grew more complex, the team moved to LangGraph to reason before acting, aggregate output from multiple nodes, and run independent branches in parallel. Realm-X can determine relevant actions, compute fallbacks, and query help pages at the same time before returning suggestions or operational actions across residents, vendors, units, bills, and work orders. LangSmith adds production visibility through traces, feedback collection, error, cost, and latency charts, plus automatic and LLM-based evaluators that monitor system health and support dynamic few-shot prompting for higher-accuracy text-to-data workflows.
A property manager asks Realm-X a question or requests an action, LangGraph fans the request into branches that identify relevant actions, compute fallbacks, and run help-page question answering, then the assistant aggregates the results and LangSmith records traces, feedback, latency, and evaluation signals for later refinement.
Evals are run in CI, with results tracked and integrated into PRs.
One major benefit of LangGraph has been its ability to run independent code branches in parallel.
performance significantly increased from ~40% to ~80%.
saving over 10 hours a week
tracking error rates, costs, and latency to keep everything on course
Review how this reference was scored and trace each normalized claim back to its documented source.
Scoring methodology
Customer Reference Value measures usefulness as a customer reference. Evidence Confidence measures the quality and traceability of available evidence, not independent verification.
Traceability
Field coverage shows which facts each source supports. Dates are shown only when documented by the source.
AppFolio lists its "Corporate headquarters" as "70 Castilian Dr., Goleta, CA 93117."
The Yahoo Finance company profile lists "Full Time Employees: 1,702."
Related
Anthropic Sample Archived · Aug 30, 2026 Atlassian Atlassian wanted agents to do consequential work inside Jira, Confluence, and related products, but enterprise customers would only trust long-running automation if the models could follow complex instructions, use… 5 million+ agents executed in customer business workflows each month Claude Code · Claude Platform · Rovo Chat Anthropic Sample Aug 21, 2026 Notion Notion needed to embed trusted Claude agents natively inside collaborative workflows, with automations, long-running tasks, memory, coding sandboxes, and human approvals that could scale across whole workspaces instead… 12 hours of prototyping work collapse into about 20 minutes Claude Managed Agents 
ServiceNow Sample Modified · Jul 30, 2026 ServiceNow ServiceNow wanted to prove GenAI value quickly for employee and customer service teams by using summarization, search, and self-service without long deployment cycles. $10M In forecast in tangible benefit in first 120 days AI Search · Knowledge Base · Now Assist 
