Cost control is product design. Every unnecessary token, retrieval query, retry, and premium-model call becomes a margin problem once every property and tenant uses the feature.
Treat this as an operating document, not a reading assignment. For Belle Realty or a property-management assistant, define the unit of work, owner, allowed failure modes, and the customer-visible consequence before choosing a model or framework. A system that gives plausible answers but cannot be measured, contained, or recovered is not production-ready.
Design stance
Start with a narrow contract. State what enters the boundary, what the component may read or change, what it must return, and when it must stop. Put the contract in version control alongside representative tenant, listing, maintenance, and leasing cases. Prefer a deterministic rule when one exists; use model judgment only where language or ambiguity genuinely adds value.
Operating loop
Instrument every request with a correlation ID, model and prompt version, retrieval version, tool calls, latency, token usage, policy decisions, and final outcome. Review a small, stratified sample of real traffic weekly. Track failures by property, workflow, language, channel, and tenant impact; aggregate averages hide the exact bad experience that produces churn.
Controls
Set an explicit threshold that changes behavior: block, require approval, degrade to search-only, route to a cheaper model, or page an owner. Test the threshold with synthetic failures and recent production examples. Do not make a dashboard metric a promise unless a named person can act on it within a defined window.
Production exercise
Apply this to a maintenance triage request: a resident reports a gas smell at 2 a.m. Document the safe path, the tool permissions, the human escalation, the message shown to the resident, and the evidence retained for review. Ship only after the behavior is repeatable in a rehearsal.
Exit criteria
The implementation has a written contract, measurable outcomes, a failure path, an accountable owner, and a rollback or containment action. If any is missing, it is still a prototype.
Cache only safe, stable work: normalized policy answers, embeddings, retrieval results with version keys, prompt prefixes, and idempotent tool responses. Never cache tenant-specific answers across authorization boundaries.
Treat this as an operating document, not a reading assignment. For Belle Realty or a property-management assistant, define the unit of work, owner, allowed failure modes, and the customer-visible consequence before choosing a model or framework. A system that gives plausible answers but cannot be measured, contained, or recovered is not production-ready.
Design stance
Start with a narrow contract. State what enters the boundary, what the component may read or change, what it must return, and when it must stop. Put the contract in version control alongside representative tenant, listing, maintenance, and leasing cases. Prefer a deterministic rule when one exists; use model judgment only where language or ambiguity genuinely adds value.
Operating loop
Instrument every request with a correlation ID, model and prompt version, retrieval version, tool calls, latency, token usage, policy decisions, and final outcome. Review a small, stratified sample of real traffic weekly. Track failures by property, workflow, language, channel, and tenant impact; aggregate averages hide the exact bad experience that produces churn.
Controls
Set an explicit threshold that changes behavior: block, require approval, degrade to search-only, route to a cheaper model, or page an owner. Test the threshold with synthetic failures and recent production examples. Do not make a dashboard metric a promise unless a named person can act on it within a defined window.
Production exercise
Apply this to a maintenance triage request: a resident reports a gas smell at 2 a.m. Document the safe path, the tool permissions, the human escalation, the message shown to the resident, and the evidence retained for review. Ship only after the behavior is repeatable in a rehearsal.
Exit criteria
The implementation has a written contract, measurable outcomes, a failure path, an accountable owner, and a rollback or containment action. If any is missing, it is still a prototype.
Measure prefill, decode, queueing, cache hit rate, tokens per request, and abandoned streams. Use smaller routers, constrained outputs, batching, prefix caching, and retrieval discipline before sacrificing customer quality.
Treat this as an operating document, not a reading assignment. For Belle Realty or a property-management assistant, define the unit of work, owner, allowed failure modes, and the customer-visible consequence before choosing a model or framework. A system that gives plausible answers but cannot be measured, contained, or recovered is not production-ready.
Design stance
Start with a narrow contract. State what enters the boundary, what the component may read or change, what it must return, and when it must stop. Put the contract in version control alongside representative tenant, listing, maintenance, and leasing cases. Prefer a deterministic rule when one exists; use model judgment only where language or ambiguity genuinely adds value.
Operating loop
Instrument every request with a correlation ID, model and prompt version, retrieval version, tool calls, latency, token usage, policy decisions, and final outcome. Review a small, stratified sample of real traffic weekly. Track failures by property, workflow, language, channel, and tenant impact; aggregate averages hide the exact bad experience that produces churn.
Controls
Set an explicit threshold that changes behavior: block, require approval, degrade to search-only, route to a cheaper model, or page an owner. Test the threshold with synthetic failures and recent production examples. Do not make a dashboard metric a promise unless a named person can act on it within a defined window.
Production exercise
Apply this to a maintenance triage request: a resident reports a gas smell at 2 a.m. Document the safe path, the tool permissions, the human escalation, the message shown to the resident, and the evidence retained for review. Ship only after the behavior is repeatable in a rehearsal.
Exit criteria
The implementation has a written contract, measurable outcomes, a failure path, an accountable owner, and a rollback or containment action. If any is missing, it is still a prototype.
Open-weight models offer deployment control and predictable capacity, but you inherit serving, safety, patching, observability, and model-quality ownership. Total cost includes the team that keeps it reliable.
Treat this as an operating document, not a reading assignment. For Belle Realty or a property-management assistant, define the unit of work, owner, allowed failure modes, and the customer-visible consequence before choosing a model or framework. A system that gives plausible answers but cannot be measured, contained, or recovered is not production-ready.
Design stance
Start with a narrow contract. State what enters the boundary, what the component may read or change, what it must return, and when it must stop. Put the contract in version control alongside representative tenant, listing, maintenance, and leasing cases. Prefer a deterministic rule when one exists; use model judgment only where language or ambiguity genuinely adds value.
Operating loop
Instrument every request with a correlation ID, model and prompt version, retrieval version, tool calls, latency, token usage, policy decisions, and final outcome. Review a small, stratified sample of real traffic weekly. Track failures by property, workflow, language, channel, and tenant impact; aggregate averages hide the exact bad experience that produces churn.
Controls
Set an explicit threshold that changes behavior: block, require approval, degrade to search-only, route to a cheaper model, or page an owner. Test the threshold with synthetic failures and recent production examples. Do not make a dashboard metric a promise unless a named person can act on it within a defined window.
Production exercise
Apply this to a maintenance triage request: a resident reports a gas smell at 2 a.m. Document the safe path, the tool permissions, the human escalation, the message shown to the resident, and the evidence retained for review. Ship only after the behavior is repeatable in a rehearsal.
Exit criteria
The implementation has a written contract, measurable outcomes, a failure path, an accountable owner, and a rollback or containment action. If any is missing, it is still a prototype.
Route by required capability, latency, geography, availability, and price—not a single leaderboard. Maintain comparable evals, data-processing terms, fallback behavior, and a migration-tested abstraction layer.
Treat this as an operating document, not a reading assignment. For Belle Realty or a property-management assistant, define the unit of work, owner, allowed failure modes, and the customer-visible consequence before choosing a model or framework. A system that gives plausible answers but cannot be measured, contained, or recovered is not production-ready.
Design stance
Start with a narrow contract. State what enters the boundary, what the component may read or change, what it must return, and when it must stop. Put the contract in version control alongside representative tenant, listing, maintenance, and leasing cases. Prefer a deterministic rule when one exists; use model judgment only where language or ambiguity genuinely adds value.
Operating loop
Instrument every request with a correlation ID, model and prompt version, retrieval version, tool calls, latency, token usage, policy decisions, and final outcome. Review a small, stratified sample of real traffic weekly. Track failures by property, workflow, language, channel, and tenant impact; aggregate averages hide the exact bad experience that produces churn.
Controls
Set an explicit threshold that changes behavior: block, require approval, degrade to search-only, route to a cheaper model, or page an owner. Test the threshold with synthetic failures and recent production examples. Do not make a dashboard metric a promise unless a named person can act on it within a defined window.
Production exercise
Apply this to a maintenance triage request: a resident reports a gas smell at 2 a.m. Document the safe path, the tool permissions, the human escalation, the message shown to the resident, and the evidence retained for review. Ship only after the behavior is repeatable in a rehearsal.
Exit criteria
The implementation has a written contract, measurable outcomes, a failure path, an accountable owner, and a rollback or containment action. If any is missing, it is still a prototype.