Skip to content

Environmental Policy

Purpose

This policy describes how RunWhen manages the environmental impact of its business operations, its cloud infrastructure, and its supply chain. It sets out the commitments we make, the practices we follow to meet them, and who is accountable for them.

RunWhen is a small, software-only company with a fully remote workforce. We do not manufacture physical goods, and we do not operate data centers or offices. Most of our environmental footprint comes from business travel, AI model inference, and cloud computing, with smaller contributions from remote work and employee equipment (see Measuring the Environmental Impact of RunWhen Operations). Water use, hazardous materials, and biodiversity impacts are not material to our operations. This policy focuses on the areas where we do have an impact.


Scope

This policy applies to:

  • All employees, contractors, and founders.
  • The RunWhen SaaS platform and the internal systems used to build, test, and operate it.
  • Company-issued hardware and business travel.
  • Selection and review of vendors and subprocessors.

Self-hosted RunWhen deployments run on infrastructure that the customer owns and operates. Those environments fall under the customer’s own environmental policies.


Commitments

RunWhen commits to:

  • Comply with all environmental laws and regulations that apply in the jurisdictions where we operate.
  • Estimate our greenhouse gas emissions using the GHG Protocol, and refine the estimate as better data becomes available.
  • Make sure employees and contractors understand this policy and their part in it.
  • Prefer infrastructure and vendors with credible, published commitments to renewable or carbon-free energy.
  • Design and operate our platform to use computing resources efficiently.
  • Minimize waste from company hardware, and dispose of retired equipment responsibly.
  • Avoid unnecessary travel.
  • Review this policy and our practices at least annually, and improve them as the company grows.

Cloud Infrastructure

The RunWhen SaaS platform is hosted on Google Cloud Platform (GCP) and runs on Google Kubernetes Engine (GKE). We do not operate our own data centers or server rooms.

  • Provider selection: Hosting on a hyperscale cloud provider uses shared, high-utilization infrastructure that is typically more energy-efficient than self-managed servers. Google publishes its environmental goals, along with region-level carbon-free energy data, in its annual environmental reports.
  • Region selection: When choosing or adding hosting regions, we consider the carbon-free energy data that Google Cloud publishes for each region, alongside customer data-residency, latency, and availability requirements.
  • Right-sizing: Workloads are sized to their actual demand, and cluster capacity scales with load so that idle resources are not left running.
  • Non-production environments: Development, test, and demo environments are shut down or scaled down when they are not in use.
  • Data minimization and retention: We collect only the data needed to operate the platform (see Data Security and Privacy), and we keep operational logs for short retention periods. Storing less data also means using less energy.

AI and Compute Usage

RunWhen uses large language models (LLMs) to support reasoning and summarization during automated troubleshooting. Model inference consumes significant compute, so we:

  • Match the model to the task, using smaller or more efficient models where they meet quality requirements.
  • Avoid unnecessary model calls.
  • Minimize the data sent to models, which also reduces the compute spent on each request.
  • Run diagnostic tasks on demand or on a defined schedule, rather than continuously.
  • Prefer LLM providers that publish environmental commitments for their infrastructure.

Workplace and Remote Work

  • RunWhen operates as a remote-first company and does not maintain dedicated office space. This removes office energy use and greatly reduces employee commuting.
  • Meetings are held by video conference by default.
  • Company records and back-office systems are run on SaaS platforms rather than on-premises servers (see Business Continuity & Resilience).
  • We prefer digital documents and electronic signatures over printed paper.

Hardware Lifecycle

  • Company-issued laptops are purchased for long service lives and are kept in use as long as they remain secure and supported.
  • Devices are repaired or reassigned before they are replaced, where practical.
  • Before reuse or disposal, devices are securely wiped in line with the End User Accounts and Device Policy.
  • Retired devices are recycled through the manufacturer’s take-back program or a certified electronics recycler. They are never discarded as general waste.

Business Travel

  • Travel happens only when meeting in person adds clear value, such as customer onsite work, team offsites, or key industry events.
  • When planning trips, employees are encouraged to combine meetings into a single trip and to choose lower-emission options, such as rail travel, where practical.

Suppliers and Procurement

  • When we select new infrastructure or SaaS vendors, we consider their published environmental commitments alongside security, privacy, cost, and capability.
  • Environmental considerations are included in the annual review of critical vendors (see Annual Security Policy Review).

Measuring the Environmental Impact of RunWhen Operations

RunWhen has carried out a preliminary assessment of the greenhouse gas (GHG) emissions from its business operations, cloud infrastructure, AI inference, business travel, and remote work.

  • Reporting period: 2026 (annualized)
  • Assessment type: Preliminary GHG emissions assessment
  • Methodology: GHG Protocol Corporate Accounting and Reporting Standard
  • Status: Internal estimate, not independently verified

Summary

RunWhen’s preliminary annualized emissions are approximately 10.1 metric tons of CO₂ equivalent (tCO₂e), with an illustrative range of 3.5–20.5 tCO₂e. All of these emissions are Scope 3 (value chain) emissions, mainly from business travel, AI inference, and cloud infrastructure.

The estimate is built from the company’s year-to-date (YTD) operating plans, supplemented by industry benchmarks and estimated emissions factors where direct consumption data is not available. It is intended to support supplier environmental compliance assessments and internal sustainability planning. It is not an independently verified carbon inventory or a certified carbon footprint, and it is not a claim of carbon neutrality.

Estimated Annual Emissions

ScopeEmissions sourceEstimated annual emissions (tCO₂e)
Scope 1Direct emissions0.0
Scope 2Purchased energy0.0
Scope 3Cloud infrastructure2.4
Scope 3AI inference and services3.0
Scope 3Business travel3.5
Scope 3Remote working0.7
Scope 3Employee equipment0.5
Total estimated annual emissions10.1

These figures are preliminary screening estimates, not directly measured emissions.

  • Scope 1 (direct emissions): RunWhen has not identified any material sources of direct emissions, such as fuel combustion, company vehicles, or on-site equipment.
  • Scope 2 (purchased energy): RunWhen has a fully remote workforce and no company-controlled offices, so it has not identified any material purchased electricity, heating, or cooling.
  • Scope 3 (value chain emissions): These are emissions from purchased services, technology infrastructure, business travel, remote work, and employee equipment. Each category is described below.

How Each Source Is Estimated

All Scope 3 estimates use assumptions built from the company’s YTD operating plans, combined with the benchmarks and emissions factors described below.

  • Cloud infrastructure (2.4 tCO₂e): Covers compute, storage, networking, and supporting data-center operations on Google Cloud Platform. This is estimated with a preliminary expenditure-based method. Where available, provider-specific data from Google Cloud’s Carbon Footprint reporting will be used to refine it.
  • AI inference and services (3.0 tCO₂e): Covers the large language models and AI services behind RunWhen’s automated infrastructure diagnostics, operational investigations, and remediation workflows. These emissions come mainly from the electricity used for model inference and supporting compute, including third-party AI services. This is estimated with indicative industry emissions factors. Actual emissions depend on model selection, inference volume, hardware efficiency, data-center location, and each provider’s energy mix. Where AI services are already included in cloud-provider emissions reporting, the overlap will be excluded to avoid double counting.
  • Business travel (3.5 tCO₂e): Covers domestic and international air travel, accommodation, and ground transportation for customer meetings, commercial engagements, and industry events. This is estimated with representative travel emissions factors. Actual emissions vary with flight distance, cabin class, travel frequency, and mode of transport.
  • Remote working (0.7 tCO₂e): Covers the additional household electricity, heating, cooling, and technology use attributable to work. Only this incremental energy use is counted, not total household energy use. This is estimated with representative home-working energy benchmarks.
  • Employee equipment (0.5 tCO₂e): Covers the embodied carbon of laptops, monitors, peripherals, and other equipment needed to support business operations. This is estimated with representative equipment lifecycle emissions benchmarks.

Uncertainty and Limitations

  • Cloud infrastructure and AI inference emissions are estimated from indirect indicators, not from measured electricity use or verified provider-specific emissions data.
  • Business travel, remote working, and equipment emissions are based on representative activity assumptions and industry benchmarks.
  • The 3.5–20.5 tCO₂e range reflects alternative estimation scenarios. It is not a statistically validated confidence interval.
  • The assessment is not a complete inventory of every Scope 3 category. Some purchased services and other indirect activities may not be included.

Reducing and Refining

The practices described in this policy are how RunWhen works to limit these emissions: shared cloud infrastructure instead of company-owned data centers, remote-first operations, efficient use of cloud and AI compute, limiting travel, and extending the useful life of equipment. The environmental benefits of these practices have not been independently quantified.

RunWhen intends to refine this assessment as more detailed activity data and cloud-provider emissions data become available.


Customer Requests

Customers and prospects can request information about our environmental practices, including responses to sustainability sections of vendor questionnaires. Contact security-and-compliance@runwhen.com.


Roles and Responsibilities

  • Company leadership approves this policy and is accountable for meeting its commitments.
  • The Security and Compliance Team maintains this policy, tracks applicable regulations and customer requirements, and coordinates the annual review.
  • The Engineering team applies the infrastructure and compute-efficiency practices described above.
  • All employees and contractors are expected to follow this policy in their day-to-day work.

Governance

This policy is owned by company leadership together with the Head of Security/Compliance. It is reviewed at least annually, or whenever our operations change significantly, such as adding hosting regions, opening an office, or reaching a size at which formal emissions reporting is required.

For any questions or clarifications regarding this policy, please contact security-and-compliance@runwhen.com.