Build an AI-Native Property Management CRM + WhatsApp AI Agent

We are a growing UK property management and lettings business looking for a strong developer, small team or technical partner to help us build our own AI-native property management platform. The objective is to create one modern platform that combines: * A property management CRM * Third-party integrations * An AI agent capable of taking actions and running workflows * WhatsApp as the main tenant communication channel * Automated property management operations This is not a basic chatbot project. We want to build a genuine operational system that could eventually replace much of the manual work involved in managing residential property. ## The Core Platform The CRM will act as the central source of truth for: * Properties and rooms/units * Landlords * Tenants * Tenancies * Contractors * Maintenance * Complaints * Documents * Communications * Tasks * Enquiries/viewings * Workflow and audit history We want a clean, modern system designed specifically around our workflows rather than trying to replicate every feature of traditional property management software. ## Integrations We do not want to rebuild specialist systems that already perform certain functions well. We intend to integrate with platforms such as: * LettsPay for rent collection * Goodlord for referencing/onboarding * InventoryBase for inventories, inspections and check-outs These should connect to our platform through APIs/webhooks where available. The integration architecture should be modular so that suppliers can be changed in future without rebuilding the wider platform. ## AI + WhatsApp Tenants will communicate with us primarily through WhatsApp. The AI agent should identify the tenant, understand their property and tenancy, retrieve relevant CRM information and handle the full range of routine tenant enquiries. Examples include: * Maintenance * Rent questions * Tenancy questions * Complaints * Housemate issues * Notices/moving out * Keys/access * General property questions * Emergency triage The AI should not simply answer questions. It should be capable of taking permitted actions, creating structured records, triggering workflows, contacting contractors, coordinating appointments, following up automatically and escalating cases requiring human approval. For example, a tenant reporting a leaking shower could potentially move from the initial WhatsApp message through diagnosis, maintenance job creation, contractor coordination, appointment booking, follow-up and completion without a staff member manually coordinating every stage. ## What We Are Looking For We want someone with strong experience in areas such as: * Production AI agents / LLM applications * Tool/function calling * Workflow orchestration * Full-stack software development * SaaS architecture * CRM/database development * API integrations and webhooks * WhatsApp Business API * Authentication and permissions * Audit trails * Reliable backend systems We are not looking for a solution that is simply ChatGPT connected to WhatsApp using Zapier/Make. We want a robust, maintainable and scalable software platform. ## Please Read the Attached Specification We have attached a much more detailed product specification covering: * CRM requirements * Integrations * AI architecture * WhatsApp workflows * Maintenance * Tenant communications * Human approval/escalation * Internal AI * Architecture * MVP approach Please review this before responding. We would particularly like your proposal to cover: 1. The most similar production system you have built 2. How you would architect this platform 3. What you would recommend for the first MVP 4. Key technical risks or concerns you see 5. How you would handle AI safety, permissions and workflow reliability 6. How you would structure the development into phases 7. Indicative cost by major component 8. Ongoing support/development model 9. Whether you are an individual developer, team or agency 10. Who would actually work on the project We are open to beginning with a paid technical discovery/architecture phase before committing to the full build. We are more interested in finding the right technical partner and receiving a realistic, well-reasoned proposal than fitting the project into a predetermined budget.