The AI copilot for building owners, facility directors, and chief engineers.
Ask anything about your property in plain English or Thai. Get answers grounded in live sensor telemetry, engineering models, and verified building graphs — backed by audit grades and human-approved work orders.
Zero-Hallucination Grounding
Every figure cites exact sensor feeds, formulas, or manuals.
Propose-Only Safety
AI proposes actions; named humans approve; zero unsupervised control writes.
Bilingual First-Class
Full natural language fluency in English and Thai (ไทย).
ISO 27001 & PDPA Compliant
3-year immutable audit log with dual-consent privacy protection.
Perplexity for your buildings
Alto Copilot is a multi-tenant enterprise intelligence platform designed specifically for commercial real estate, hotels, hospitals, retail malls, and industrial facilities.
While generic AI tools search the public internet, Alto Copilot connects directly to your property’s internal systems — BMS/BAS controllers, IoT meters, chilled water plants, AHUs, digital twin ontology graphs, standard operating procedures (SOPs), and equipment warranties.
When anyone from the boardroom to the plant room asks a question, Alto Copilot retrieves the exact telemetry, applies certified energy-engineering methods, and produces an answer with a transparent Claim Grade and an actionable proposal.
Four Architectural Pillars of Trust
01
Grounded & Graded Answers (No Hallucinations)
The Problem
Generic language models make up numbers, creating dangerous risks for multi-million-dollar HVAC plants.
The Alto Solution
Every metric returned by Alto Copilot carries a Method Note and a Claim Grade (measured, derived, estimated, modelled). Users can click any number to inspect the formula, data source, and exact timestamp.
02
Propose-Only Safety Invariant (Invariant D9)
The Problem
Handing automated control of chiller plants to an autonomous LLM is an unacceptable operational risk.
The Alto Solution
Alto Copilot never writes directly to physical control systems. It generates an ActionProposal with estimated energy savings and operational risks. A qualified human with approval rights must sign off, generating an audited WorkOrder for dispatch in your existing CMMS or BAS.
03
Native Bilingual Fluency (English & Thai)
The Problem
Corporate executives communicate in English, while field technicians and facility operators work in Thai.
The Alto Solution
Alto Copilot natively processes technical HVAC terminology, SOP queries, and refusal reasons in both languages. A refusal displays as “Requires Engineering Manager approval” on the corporate dashboard and “รออนุมัติหัวหน้าช่าง” on the technician’s mobile interface.
04
Sovereign & Adaptive Model Routing (Invariant D3)
The Problem
Real estate portfolios have strict data residency, privacy, and budget constraints.
The Alto Solution
Queries route through a multi-tier policy engine: Sovereign Tier (on-premise / private infrastructure such as DGX) for confidential telemetry; Edge Tier (local gateways) for latency-sensitive tasks; Cloud Tier (L1–L4) with smart cost-routing and a $0 preview inspection engine.
The Specialized Multi-Agent Intelligence Swarm
Alto Copilot orchestrates specialized AI agents built specifically for building engineering.
HVAC Expert
hvac_expert
Diagnoses chiller staging, condenser water delta-T syndrome, AHU/FCU psychrometrics, and sensor drift. Detects faults before equipment breaks.
Measurement & Verification Analyst
mv_analyst
Applies IPMVP Option C standards to measure normalized energy savings against weather, occupancy, and historical baselines.
Facility Operations Specialist
facility_ops
Instantly retrieves building SOPs, emergency manuals, warranty clauses, and maintenance histories for on-site staff.
Sustainability & ESG Reporter
sustainability_reporter
Tracks Scope 1 & 2 carbon emissions, energy use intensity (EUI), and generates audit-ready ESG sustainability reports.
General Assistant with Live Web Search
general_assistant
Answers general inquiries and conducts real-time web research on local utility tariffs, weather alerts, and supplier specs via MCP.
Code Studio Assistant
coding_agent
Assists facility data teams with custom telemetry scripts, automation routines, and data transformations.
Standardized MCP (Model Context Protocol) Connectors
Alto Copilot connects directly to facility infrastructure through open, propose-only MCP connectors.
alto-ontology
Navigates the digital twin relationship graph (spaces, floors, chillers, pumps, air loops).
alto-timeseries
Queries interval sensor telemetry, power meters, enthalpy, and thermal loads.
alto-knowledge
Hybrid BM25 and vector semantic search across manuals, blueprints, and operational docs.
alto-workorder
Issues structured, traceable work orders without manual data entry.
alto-websearch
Real-time external knowledge integration via TinyFish MCP server.
Human-in-the-Loop Approvals & Work Order Dispatch
Dedicated Approvals Queue (/approvals) for authorized managers to review proposed setpoint adjustments, maintenance alerts, or scheduling shifts.
Full impact analysis: estimated kWh savings, monetary impact (THB/USD), and risk level displayed prior to human approval.
One-click approval logs the decision into an immutable audit stream and dispatches the work order.
Multi-Tenant Workspaces & Privacy-Safe Memory
Projects Workspaces (/projects): isolate conversations, custom agent prompts, and sensitive documents across buildings or departments.
Semantic Memory with Dual-Consent PDPA: remembers tenant preferences and historical operational patterns, protected by mandatory organizational and personal consent gates, with automatic 90-day episodic memory expiry.
Enterprise Auditability & ISO 27001 Posture
Append-Only Activity Log (/activity): 3-year tamper-proof audit trail capturing every user prompt, agent step, model tier, and human action.
Google ADK EvalOps Framework (/evals): continuous automated 3-tier testing pyramid ensuring response groundedness, safety rubric compliance (“The Law”), and trajectory alignment.
Interactive 3D Digital Campus Experience
A live 3D isometric simulation connecting 7 real-world facility roles across an office tower, hotel, retail mall, and central cooling plant (/experience).
Experience firsthand how an incident (e.g., AHU cooling loss) flows from detection to diagnosis, proposal, approval, and resolution.
From question to verified action
Every insight is grounded, every action is proposed, and every approval is made by a named human.
Ask Alto Copilot
How much energy did we save against baseline at Sukhumvit Grand this month?
Why is Chiller #2 running at low delta-T during peak hours?
AHU-B3 เสียงดังและไม่ทำความเย็น ต้องตรวจสอบตามขั้นตอนอย่างไร?
Intent Classification
Agent Selection · hvac_expert
Tool Calls · alto-timeseries
Grounded Answer · measured
01
Copilot Analyzes & Proposes
Suggests resetting chilled water supply temperature by +0.8°C (Estimated savings: 4.2% chiller power).
02
Human Approval Gate
Chief Engineer reviews risk assessment and clicks “Approve”.
03
Work Order Dispatched
Generates WO-0842 for execution. No direct BMS write-back.
The Grounding Inspector
14,280 kWh · ฿61,400
Click any figure to inspect exactly how it was calculated.
Method
IPMVP Option C Whole Facility Regression
Inputs
15-minute interval meters, weather normalization (CDD)
Grade
Derived · Verified
Citation
Baseline Model Ref: BL-2026-Q3
Who benefits from Alto Copilot?
Frequently Asked Questions
Does Alto Copilot directly control or change settings in our Building Management System (BMS)?
No. By design (Invariant D9: Propose-Only), Alto Copilot never directly actuates physical control systems. It analyzes telemetry, identifies optimization opportunities, and drafts structured Action Proposals. A qualified human must review and approve the proposal, which then issues an audited Work Order for your maintenance team or BMS operator.
Can our field technicians use Alto Copilot in Thai?
Yes. Thai and English are both first-class languages in Alto Copilot. The system features a custom bilingual engineering lexicon that understands Thai HVAC terms, local equipment slang, and translates standard operating procedures seamlessly.
How do you guarantee the AI doesn’t hallucinate numbers?
Alto Copilot uses a deterministic grounding engine. Every metric displayed is accompanied by a Claim Grade (measured, derived, estimated, or modelled) and a Method Note detailing the mathematical formula, baseline comparison, or sensor readings used. If data is missing, the system explicitly reports data unavailability rather than guessing.
How is our building data secured and kept private?
Alto Copilot is built to ISO 27001 standards and complies with Thailand’s Personal Data Protection Act (PDPA). It features strict multi-tenant isolation, role-based access control (RBAC), automatic 90-day episodic memory expiry, and an append-only, tamper-proof audit log kept for 3 years. We also support on-premise sovereign model deployments for sensitive data.
What systems and sensors can Alto Copilot connect to?
Alto Copilot connects via open Model Context Protocol (MCP) servers to BACnet/IP gateways, Modbus power meters, IoT environmental sensors, chiller plant management systems (CPMS), and CMMS work order platforms.
Ready to turn building telemetry into verified action?
Join leading hospitality groups, commercial developers, and property managers optimizing energy performance with Alto Copilot.