HVAC Comfort & Optimization for bank-branch networks: three accumulative tiers (Essential / Managed / Optimized), a Data Enabler that leads with connecting to the BMS your network already runs, product-specific add-ons, platform-wide add-ons, and a Custom advisory layer. Built for retail banking where branch comfort is part of the experience, HVAC is a large controllable operating cost replicated across a uniform estate, and every branch runs the same network-standard controls.
Each product has its own Essential / Managed / Optimized structure, its own Data Enabler, and its own set of add-ons. This document covers HVAC Comfort & Optimization, the anchor product for the segment. Lighting Control & Optimization and Energy Intelligence are now drafted to the same model. Platform-wide add-ons and the Custom layer apply across every product.
Asset class: RTUs / packaged & split units serving the branch — teller floor, back-of-house, and ATM vestibule.
Pain points: Customer and staff comfort, HVAC energy spend across the network, equipment life, comfort complaints.
Asset class: Interior, exterior / signage, and ATM-vestibule lighting.
Pain points: Lighting energy spend, scheduling, 24/7 ATM-vestibule availability.
Asset class: Whole-branch energy — metering, tariff / demand analysis, cross-site benchmarking.
Pain points: Portfolio-wide visibility and energy savings across a large, uniform network.
Impact: Up to 25–40% HVAC energy savings, comfort held in range across the branch, and 10–20% longer RTU life at the Optimized tier — with measurable wins already at Managed. Per-tier ranges in the Expected outcome row below; bank branches have no named Keedian reference deployment yet, so every figure is expected (pending validation).
What it solves: Comfort complaints from customers and staff the branch team can't diagnose remotely, runaway HVAC energy spend replicated across a large network, premature RTU failures, setpoint drift across the estate, and managing HVAC across hundreds or thousands of near-identical branches from one place.
These are the operational metrics the product monitors. What changes across tiers is who acts on them — you (Essential), Keedian's Agentic Operations (Managed), or Keedian plus continuous tuning (Optimized) — and, where the data layer adds instrumentation, which signals become available.
The core metrics are summarized below. Full methodology, thresholds, diagnosis steps, and escalation criteria are in the Metrics Catalog.
Average compressor hours per day while actively running. The baseline signal for operational health — flags equipment running too long (degraded efficiency or capacity issues) and too short (scheduling or data problems) before the pattern becomes a larger issue.
Full detail → Metrics CatalogTemperature difference a cooling stage develops across its heat-transfer surface — return vs supply air on an RTU or AHU, chilled-water return vs supply across a chiller's evaporator, or entering vs leaving water on a cooling tower. Measures how effectively the equipment moves heat. A declining or persistently low ΔT is the earliest remote signal of degradation — fouled coils or tubes, low flow, or a failing charge — often visible weeks before a comfort failure or breakdown.
Full detail → Metrics CatalogPercentage of time the thermostat setpoint matched the agreed corporate policy. Measures whether the system is being operated as configured — independent of whether the physical outcome was reached. Most reliable signal of override patterns at the site level.
Full detail → Metrics CatalogPercentage of time actual space temperature was within the expected range during evaluated hours. Measures the physical outcome — whether the equipment is delivering the required comfort. Correct setpoints with poor temperature compliance is the clearest signal of an equipment or capacity issue.
Full detail → Metrics CatalogEach metric includes definition, detection logic, diagnosis steps, recommended actions, escalation criteria, and prevention controls. Thresholds are reference values to be calibrated per client.
View Metrics Catalog →Each tier adds capabilities on top of the previous one. You never lose functionality by upgrading, and you can start at any tier.
Add-ons activate on top of any tier and any Data Enabler model. Each is either Domain-specific (within a product domain) or Transversal (across any contracted product). Open an add-on for its hardware, value levers, and pricing — or browse the full add-ons catalog.
Adjacent capabilities within the product domain. Hardware and pricing live on each add-on's page.
| Add-on | What it solves |
|---|
Transversal add-ons that apply across any product you contract, charged once per branch.
| Add-on | What it solves |
|---|---|
| BMS / EMS Integration Extension | Extend the integration to additional systems not included in the product's base Model B. |
| Custom CMMS Integration | Deep integration with your CMMS beyond the standard. |
| Data as a Service | Programmatic access to Keedian's normalized operational data for your analytics, BI, or data lake. |
| Ad-Hoc Dashboards & Reporting | Custom dashboards and one-off reports built to your spec, on top of the data Keedian already collects. |
| UPS & Continuity Monitoring | Monitoring of the electrical-continuity chain — UPS and standby generators, with the ATS and main switchboard as supporting signals — so a backup that is not ready surfaces before the outage it was meant to cover, not during it, and the site stays online: an office building's tenants and critical rooms, or a bank branch's ATMs, teller and core-banking systems, security, and comms. Monitors readiness ("will the backup take the load, and for how long?"), not just "broken now." Continuity links to the other monitoring add-ons for the layers they already cover — Chiller Plant Monitoring (critical-environment cooling), Fire Detection & Suppression Monitoring, and Vertical Transport Monitoring — which read alongside continuity but stay separate add-ons. |
Tailored work for needs outside the standard packaging — scoped and priced per engagement, on top of any product and any tier. Keedian owns the connectivity and data layer; the physical asset stays with you and your vendors. Browse the full advisory catalog.
| Advisory element | What it solves | Type | Domain |
|---|---|---|---|
| Ad-hoc Solution Definition | Scoping and definition of ad-hoc solutions for needs not described in the standard tier model. | Transversal | All products |
| Field Services | On-site work on the connectivity layer — sensor calibration, gateway and smart-thermostat installation, integration validation — plus coordination of your HVAC vendor for asset repairs. | Transversal | All products |
| Sustainment | Post-deployment maintenance of the Keedian layer — sensor recalibration, firmware updates, hardware replacement, and periodic integration validation. | Transversal | All products |
| Customer-specific SOPs | Bespoke operating procedures — vendor-specific dispatch rules, regional escalation matrices, pre-dispatch approval workflows, and brand-aligned site communication scripts. | Transversal | All products |
| Custom Reporting | Bespoke executive and operational reports built to your specification on top of the data Keedian already collects. | Transversal | All products |
| One-off Integrations | Bespoke integrations with proprietary systems or clouds for which Keedian has no standard connector. | Transversal | All products |
| HVAC Retrofit — Connectivity & Savings Evaluation | Connectivity and control-sequence design for HVAC retrofits, plus savings evaluation. Keedian scopes the controls and telemetry and quantifies expected savings — not the mechanical retrofit. | Domain-specific | hvac |
| RTU Replacement Input | A data-driven shortlist of the RTUs that are candidates for replacement, based on runtime, failure history, and efficiency signals. | Domain-specific | hvac |
| Chiller Plant Engineering — Sequence & Savings Evaluation | Sequence-of-operations design and savings evaluation for the central cooling plant. Keedian designs the chilled-water / condenser-water control sequences and quantifies expected plant savings — not the mechanical plant work. | Domain-specific | hvac |
Pricing: project-based · scoped per engagement
The Managed and Optimized tiers are delivered through Agentic Operations — AI agents and human operators working together against the same operational rules and the same SLA.
AI agents handle the routine, high-volume work. They monitor alarms in real time across the network, triage common HVAC false positives like short cycling and sensor drift, run fault detection on branch RTUs (compressor degradation, fan failures, refrigerant loss), generate setpoint and schedule recommendations driven by weather forecast and branch hours, identify outlier branches, draft work orders with the suspected root cause, and verify closure with telemetry evidence. They work 24/7 against the comfort thresholds you set.
Human operators handle the exceptions. When a comfort complaint requires conversation with the branch manager, when a setpoint change needs customer approval before acting, when a vendor escalation needs a human voice on the phone — that's the human side of the team. Operators also manage the customer relationship, refine the operating rules over time, and own the outcome reporting.
One service, one SLA, one accountability. The split between AI and human is invisible to you — you call Keedian, you read Keedian's reports, you measure Keedian against the contract. The internal mechanism evolves as the agentic stack matures; your contract does not change with it.
The Data Enabler is how the operational signal enters Keedian. Bank branches are different from a single small-retail store: corporate real estate standardizes controls across the whole network, so the default path is Model B — we connect to the BMS / EMS you already run and read its RTU points. We don't charge you to integrate it; we use the investment the network already made. Where a branch has no usable BMS — typically older or recently acquired branches on standalone thermostats — Model A installs Keedian hardware. Where coverage is partial, Model C fills the gaps. The managed service, dashboards, SLA, and outcomes are identical across all three models.
Your branch controls expose data through open protocols: BACnet, Modbus, MQTT, documented REST APIs. Tridium/Niagara, Distech, modern smart-thermostat platforms with documented APIs. This is the default branch-network path.
Your BMS / EMS / thermostat platform uses closed protocols without modern APIs and requires reverse engineering: legacy EMS, proprietary thermostat clouds without API access, email-based reporting.
The branch has no BMS / EMS or smart thermostats — or what it has can't be read remotely. Common in older or recently acquired branches on standalone controls.
You have partial coverage: a network standard BMS on most branches but standalone thermostats on the acquired ones, or a BMS that covers the main RTUs but not the ATM-vestibule unit.
Maintained by Product + Engineering. Lets Sales quote on the spot without engineering scoping. Classification is pre-assigned per system; any system not on this list requires Engineering review before quoting.
Bank branches are the hybrid case. The equipment is retail: packaged rooftop and split units serving a small single-story space, so the asset classes and the priced unit mirror the grocery / convenience / small-retail HVAC products — per RTU, per branch, not per office zone. But the connectivity is office-like: corporate real estate standardizes a BMS / EMS across the network, so the Data Enabler leads with Model B — connect to the controls you already run — the way the offices products do. Retail asset classes with a Model-B-first Data Enabler is the shape that defines the segment.
Unlike a single small-retail store, a bank branch belongs to a network whose corporate real estate function standardizes controls — a BMS, an EMS, or a smart-thermostat platform deployed across hundreds or thousands of branches. The Data Enabler therefore leads with Model B: we connect to the system you already paid for and read its RTU points. Model A — Keedian hardware end-to-end — is the exception, reserved for older or recently acquired branches whose controls can't be read remotely.
A branch is a small-format single-story space with a handful of RTUs — a teller floor, back-of-house, and often an ATM vestibule. Variable pricing is per RTU, the same retail unit the grocery and convenience HVAC products use, because RTU count varies little across branches and scales fairly with both customer value and Keedian effort.
A branch is a customer-facing retail space as much as a workplace. Comfort in the teller and waiting areas is part of the experience, and comfort for branch staff is a retention and productivity lever — so customer_experience is a first-class value driver alongside energy and equipment life, exactly as in the retail HVAC products.
Banking networks run from a hundred to several thousand near-identical branches, so an HVAC energy improvement captured once replicates across the whole estate. The Optimized tier's value proposition is heavily energy-focused — real-time, tariff-aware optimization — with reliability and comfort as the secondary benefits, the same emphasis as the retail HVAC Optimized tier.
This is a preliminary build so Sales can position Keedian for banking before final pricing and validation. Pricing is a placeholder (structure only, amounts pending maintainer input) and there is no named reference customer — the validated Keedian benchmarks are retail, not banking, so every outcome figure on this page is expected (pending validation). Real bank-branch pricing and a validated reference are separate future work.
The Essential / Managed / Optimized tiers, the Data Enabler model selection, the split between AI agents and human operators, and the SLA structure are identical to the retail HVAC products. The differences are only the Model-B-first Data Enabler (the network runs a BMS) and the branch-specific framing (a customer-facing retail space in a large uniform network).