Climative builds low carbon plans
Climative ACM automated carbon model article header

Automated Carbon Model (ACM™) for Homes: What Is It?

The Automated Carbon Model (ACM) is Climative’s powerful machine learning algorithm used to quickly and cost-effectively evaluate the energy efficiency and carbon footprint of homes. The traditional approach to home energy assessments only covers 1-3% of the building stock annually, with a post-audit retrofit rate of about 0.7%. At this rate, it will be beyond 2150 before we evaluate and retrofit all homes to net zero. The Automated Carbon Model was created to greatly accelerate energy assessments and drive impactful home retrofits. The ACM‘s high-quality home carbon data output has many uses to help banks, governments, utilities, and homeowners achieve their net zero goals. 

Quick Facts

What is the ACM?

The Automated Carbon model is Climative’s algorithm for quickly and cost-effectively evaluating homes’ energy efficiency and carbon footprint on a large scale. 

How Accurate Is it?

Trained on 1M+ on-site energy assessments, results are about 80% accurate. This estimate requires fewer than ten inputs, compared to 200+ for an on-site assessment.

What Is It Used For?

These energy and emission insights are useful for carbon reporting, net-zero policy, efficiency program design, homeowner engagement, and building labeling.

What is the Automated Carbon Model (ACM)?

The Automated Carbon Model (ACM) is Climative’s advanced algorithm for evaluating the energy efficiency and carbon footprint of residential properties on a large scale. The ACM powers Insights and Navigator, components of the Climative Platform.

What is the Automated Carbon Model Used For?

The Automated Carbon Model (ACM) is used to quickly generate valuable emissions insights for buildings, such as estimated annual energy consumption, estimated greenhouse gas emissions, a breakdown of energy usage, recommendations for energy efficiency upgrades including potential cost savings, projected GHG emission and energy efficiency ratings for each assessed home.  

Aggregate insights generated by the ACM are well-suited for policy design, efficiency program design, carbon reporting, building labeling, and more. 

For individual homeowners, the ACM produces a “first pass” assessment based on available data. The assessment can be made more reflective of the building, thus more accurate and actionable, when a homeowner adds more info in the questionnaire, linked billing data, and attached results from an on-site assessment. Climative suggests that homeowners speak to an energy efficiency professional before investing in any efficiency retrofits. 

Download Scaling Net Zero Retrofits or book a call to learn about use cases. 

How is the Automated Carbon Model Trained?

The ACM is trained using over a million on-site energy assessments conducted by registered energy advisors. The analysis can consider utility bills and census data, ensuring that the AI models are accurate compared to on-site energy assessments. The model also considers geographical and demographic factors when providing energy retrofit recommendations. 

What are the Data Inputs and Outputs of the Automated Carbon Model?

The ACM generates energy insights with very few inputs: fewer than 10, compared to over 200 collected during a typical on-site assessment. Public data (such as property assessment, energy costs, and weather data) can be used to generate a comprehensive low-carbon plan. Layering in additional data from surveys, remote assessments, and on-site assessments can further improve the accuracy and helpfulness of each low-carbon plan 

How does the Automated Carbon Model work?

The Automated Carbon Model (ACM) assesses residential buildings’ energy efficiency and carbon emissions across regions without requiring an on-site visit. It considers building characteristics, geographical location, weather impact, and detailed information about the building, such as furnace age and building envelope insulation, obtained through homeowner surveys, remote assessments, and even on-site assessments.  

Using building science-based models and machine learning techniques, the ACM is trained and tested on on-site assessments conducted by registered energy advisors.   

The output of the ACM (called a home energy report, home energy assessment, or low-carbon plan) helps homeowners better understand their houses in terms of energy efficiency and carbon emissions. Furthermore, it provides homeowners with customized and actionable energy retrofit recommendations to speed up their transition to net zero. 

Is the Automated Carbon Model Accurate?

Yes. Accuracy is about 80% compared to modeling software used by registered energy advisors when they create labels for homes. This estimate requires less than ten inputs including, but not limited to, year of construction, type of house, square footage, primary heating fuel type, and location. 

This level of accuracy makes the ACM best suited for tasks such as energy efficiency program design and carbon reporting such as that for financed emissions. Additional data (from homeowner questionnaires, remote assessments, and other sources) can be layered for more comprehensive insights suited for energy efficiency program delivery and driving retrofit actions.  

To learn more about the accuracy of the ACM, read the results of the Remote Assessments Pilot or book a call with a team member. 

Does the Automated Carbon Model Assist with Scope 3 Emissions Reporting?

Yes, the ACM assists with Scope 3 carbon emissions reporting: indirect greenhouse gas emissions that occur in a company’s value chain but are not directly controlled by the company itself. Climative’s ACM outputs can align with the latest standards, such as PCAF, to ensure that the assessment aligns with internationally recognized carbon accounting and reporting practices, enhancing its credibility and comparability for different use cases. Climative is currently validating with our partners that the home carbon rating generated with Climative’s ACM is equivalent to PCAF score 3 (reflecting a high level of understanding of emissions), while comparable methods yield a lower PCAF score 4 or 5. With the cooperation of utilities, Climative’s outputs combined with energy use data could even achieve PCAF score 1. 

How Does Climative’s Automated Carbon Model Compare to Physics-Based Approaches to Estimate Home Carbon Scores?

Professionals commonly use physics-based energy simulation modeling software to assess home energy efficiency and carbon emissions. Energy simulation modeling software typically requires an on-site visit, a blower door test, and over 200 input data fields for simulations. The slow and manual nature of this process limits the capacity to assess only 3% percent of the residential building stock each year.  

Compared to physics-based approaches, the Automated Carbon Model greatly accelerates the pace and affordability of home energy assessments. Climative’s ACM can produce market-wide assessments of over a million homes per day at 100% coverage of the building stock.  

Homeowner showing digital energy assessment to contractor

How Does Climative’s Automated Carbon Model Compare to Archetypal Analysis?

Archetypal analysis is a common approach to assessing residential building energy efficiency and carbon emissions. This approach categorizes the building inventory into broad groups. It creates an archetype home representing all buildings in the same group, which may lead to less precise estimations due to the overly broad categorization. In contrast, Climative’s ACM, which leverages machine learning, has demonstrated higher prediction accuracy than archetypal analysis, accounting for nuances that archetypal analysis cannot support.  

Climative is validating with partners that the carbon score generated with the ACM is equivalent to PCAF score 3, while archetypal analysis yields a carbon label with lower data quality (PCAF score 4 or 5). 

How Does Climative’s ACM Assessment Compare to Operational Assessments?

Climative’s ACM can combine data from operational and asset assessments to create a comprehensive digital twin of a building. 

When assessing the energy use of a building, typically, it can be divided into two categories: 

  • Operational: evaluation of a building’s energy based on its actual energy usage, rather than the construction and features. Examples of input data: number of occupants, thermostat settings, how often occupants are home, and energy bills. 
  • Asset-based: evaluation of a building’s energy efficiency based on its construction and features. Allows homes to be compared “apples to apples” regardless of how the occupants operate the building. Examples of input data: type of heating system, insulation, number of windows, and presence of on-site renewable energy generation. 

Most residential energy rating systems (such as HERS and DoE’s Home Energy Score) use an asset-based approach, and Climative uses an asset-first approach. Climative is set apart by the ability to input operational data in the form of energy bills, to enhance the home energy assessment to account for 100% of a building’s energy use. 

Does the Automated Carbon Model Produce Ethical and Equitable Results?

Climative takes significant measures to ensure our models produce energy insights that create prosperity and opportunity for all, particularly groups who have been historically underserved. We employ a rigorous Responsible AI Framework for producing virtual home energy assessments. This framework includes a comprehensive AI process to manage the risks associated with energy technology. We invite you to review our Responsible AI Framework.    

Climative responsible AI process and risk management

How Does Climative Protect the Personal Information of Homeowners?

When Climative’s ACM generates energy insights for a region, these insights are sometimes shared with the public and certain third parties. The information used to generate these insights is publicly available, such as tax assessment data, energy cost data, and weather data, and is not tied to personal identifiable information.  When homeowners use Energy Navigator to add information to their Home Energy Report (in the form of a questionnaire or linked utility bill, for example), this information is never shared without consent. 

Some of the stakeholders that are interested in these energy insights include: 

  • Governments, to inform net-zero policy 
  • Research institutions, to learn about ways to accelerate decarbonization  
  • Efficiency organizations, to design more impactful energy efficiency programs 
  • Real estate professionals, for building labeling 
  • Banks, to report on financed carbon 

Climative uses the latest web security standards and protocols to protect user data. Visit climative.ai to review our privacy policy and Responsible AI Framework. 

What are the Use Cases for the Automated Carbon Model?

With high-quality home energy data at scale, the possibilities are endless. Climative’s ACM is the backbone of our data platform (Energy Insights) and our homeowner engagement platform (Energy Navigator) to help our partners gather and understand the data required to achieve their net-zero objectives:  

  • Low carbon program design for cities, states, and provinces;  
  • Inclusive climate policy design;  
  • Program deployment and citizen engagement;  
  • Carbon reporting for banks and insurers;  
  • Home energy labeling at the time of sale;  
  • Loan origination;  
  • and more.  

Book a call or download our whitepaper, Scaling Net-Zero Retrofits, to learn more about how the Automated Carbon Model can help you accomplish your net-zero goals. 

Automated Carbon Model ™ and ACM ™ are trademarks of Climative. 

Tianshu Huang

Tianshu Huang

Tian is a dedicated and creative statistical modeler who leads the development of Climative’s machine learning algorithms. She believes that ethical and thoughtful use of AI will help us achieve climate goals while creating equal opportunity for all.

LinkedIn
Facebook
Twitter
LinkedIn

Interested in Changing the trajectory of climate change?

Climative is building a climate-positive world through innovative, collaborative, and data-driven solutions.

This is your invitation to be part of the journey

Low Carbon Dashboard