Rethinking CRM for an AI-Native World
a real estate CRM that understands work, surfaces intent, and drives outcomes
problem
CRMs are built to store interactions, but selling depends on understanding them. Conversations happen across meetings, Slack threads, emails, and informal discussions, while the system waits for users to update and interpret it. The software records what happened after the fact instead of helping decide what should happen next. Users must notice risk, decide next steps, and manually drive progress.
solution
The product is reframed as an operational partner, not a database. An AI assistant watches activity in real time, proactively identifies meaningful moments, and opens a conversational space to investigate context and confirm decisions. Once approved, actions are executed automatically. Instead of managing pipelines, users supervise outcomes while the assistant continuously runs the workflow.
The project explores the human control-plane for an AI-native real estate CRM where agents no longer manage pipelines manually, but supervise decisions.
Progress happens through conversations, tours, questions about pricing, negotiation, document exchange, and renewal discussions. Each interaction changes confidence, risk, and intent long before a deal is marked won or lost. Most of this context lives outside the CRM: calls, meetings, WhatsApp messages, emails, and property visits. The following flow illustrates how leasing actually progresses: relationships evolve through interactions, signals accumulate over time, and decisions emerge from context rather than stages.
To meaningfully assist and shift the CRM from a passive record to an active participant in the workflow, the system must:
Act proactively to surface signals, risks, and next steps without waiting for input
Be conversational to allow users to explore context, validate reasoning, and make decisions in natural language
year
2026
timeframe
48 hours
tools
Figma, ChatGPT, Framer
category
Product design
see also
