challenge-image

AI-Powered Building Energy Ratings: Faster, Cheaper, and Accessible to All

Short video and F6S Link

This challenge invites you to build an AI-native application that dramatically accelerates BER assessments by using image recognition, structured data capture, or publicly available property datasets to estimate a building’s BER grade (A–G) and generate improvement recommendations without a full manual survey. The required deliverable is a working prototype supported by technical documentation. This challenge is designed to give you real-world business experience that matters, whether you are building your portfolio, exploring entrepreneurship, or launching a career in AI.

(By registering for this challenge, participants agree to abide by the Non-Disclosure Agreement (NDA) governing access to EnergyBright's proprietary data and methodology materials, a copy of which is available in the Resources section.)

INCENTIVES

Equity participation: A stake in a joint venture or in EnergyBright's AI product line, reflecting your contribution to building the technology. IP licensing: If you prefer to retain ownership of your technology, a licensing arrangement where EnergyBright pays to use your solution commercially. Revenue sharing: A model where you earn a percentage of revenue generated by assessments conducted through the platform.

Service contract: A paid development engagement to take the MVP to production quality, with ongoing maintenance and support.

Hybrid arrangements: Any combination of the above, structured to align incentives for both parties.

Detailed Problem Statement

Every building in Ireland must have a Building Energy Rating (BER) before it can be sold or rented. The same obligation exists across the EU under the Energy Performance of Buildings Directive. Yet the process remains stubbornly manual: a certified assessor visits every property, surveys every room, records data on insulation, heating systems, windows, ventilation, and construction materials, and then enters it all into specialized software to generate a rating on an A-to-G scale. A single assessment takes between 30 minutes and over an hour on-site, before accounting for travel and administration.

The result is a bottleneck. Homeowners face delays and costs. Large-scale retrofit programs- the kind Europe urgently needs to meet its climate targets- cannot scale fast enough. The housing market and the energy transition are both being held back by a process that has not fundamentally changed in decades.

The challenge runs deeper than inconvenience. Certified assessors are a limited resource. Site visits cannot be parallelized. The economies of scale that digital tools typically deliver simply do not apply when every assessment demands a physical presence.

Meanwhile, the revised EPBD (2024) is pushing explicitly toward greater digitalization and harmonization of energy certificates across Europe- a clear policy signal that the moment for an AI-native solution has arrived.

The question is straightforward: can AI make building energy assessments dramatically faster, cheaper, and more accessible without sacrificing accuracy?

EnergyBright believes the answer is yes. This challenge is the invitation to prove it.

Root causes
  • Manual assessments require a certified assessor to physically visit every property, making the process impossible to scale.
  • The pool of certified assessors is finite and cannot expand fast enough to meet growing retrofit demand across Europe.
  • The building elements that matter most for energy performance- insulation, wall composition, and roof construction- are hidden and rarely documented.
  • Many homes have been upgraded with new boilers, insulation, or windows, but the paperwork to prove it is frequently missing or incomplete.
  • No AI-native tool currently exists that can approximate a BER assessment faster, cheaper, and through an interface non-technical users can actually use.
Previous Attempts to Solve
  • Existing BER software tools such as Ireland's DEAP digitized the calculation engine but left the data collection process entirely manual and assessor-dependent.
  • Remote or desktop-based assessment approaches have been explored but struggle with accuracy due to incomplete building data and lack of physical verification.
  • Age-based default values were introduced to speed up assessments where documentation is missing, but they produce conservative estimates rather than accurate, building-specific ratings.
  • Computer vision and satellite imagery have been applied to building stock analysis at a research level, but no production-ready tool has successfully translated this into reliable individual property ratings.
  • European governments have pushed toward Energy Performance Certificate (EPC) harmonization and digitalization through the revised Energy Performance of Buildings Directive (EPBD), but policy ambition has outpaced the availability of practical technology to deliver it.
Urgency and Importance

Europe's climate targets demand a massive acceleration in building retrofits, but retrofits cannot happen without energy assessments, and assessments cannot scale under the current manual model. Every week of delay means more homes sold or rented without reliable energy data, more retrofit programs bottlenecked, and more emissions unaddressed. The revised EPBD (2024) has made digitalization of energy certificates a policy priority across all EU member states. The window to build the defining tool in this space is open now — and it will not stay open indefinitely.

Target Beneficiaries
  • Homeowners seeking a faster, cheaper way to understand their property's energy performance before selling, renting, or planning retrofits.
  • Real estate agents who need energy ratings quickly to keep property transactions moving without delays.
  • Energy advisors and retrofit programs that require assessments at scale to identify and prioritize buildings for upgrading.
  • Housing authorities and large landlords managing property portfolios who need efficient, batch-level energy performance data.
  • Governments and policymakers across Europe are working to accelerate building decarbonization and meet EPBD compliance obligations.
  • Renters and buyers who deserve clear, accessible energy information to make informed decisions about the homes they occupy.

Objectives
  • Build a working AI-native MVP that can capture building data and produce an accurate Irish BER grade and energy performance score in significantly less time than a manual assessment.
  • Demonstrate that AI can approximate the output of Ireland's DEAP calculation methodology using inputs that are faster and easier to collect than a full manual survey.
  • Deliver an experience accessible to non-technical users — including homeowners, real estate agents, and energy advisors- that works on mobile devices.
  • Design a modular architecture from the outset that can accommodate different national EPC calculation methodologies beyond the Irish BER system.
  • Produce transparent, actionable outputs that clearly show the estimated rating, the key factors driving it, the confidence level, and recommended improvements.
Desired Outcomes
  • A functional MVP that can predict the correct Irish BER grade- or come within one grade- for a representative sample of buildings provided by EnergyBright.
  • An AI assessment pathway that demonstrably reduces on-site data collection time compared to the current manual process of up to one hour per property.
  • A user interface validated as usable by non-technical users, accessible on mobile devices, and capable of producing clear rating outputs with confidence levels and improvement recommendations.
  • An architecture document demonstrating how the system could be extended to support at least one additional European EPC framework beyond the Irish BER.
  • A validated commercial pathway, with the winning solution forming the foundation of a real product through an equity, licensing, revenue-sharing, or service arrangement with EnergyBright.

Stretch goals (not required, but valued):

  • Modular architecture that allows swapping in different national calculation methodologies
  • Ability to process floor plans or building blueprints
  • Integration with public data sources (e.g., property registries, geospatial data, weather data)
  • Comparison mode showing how a building would rate under different European EPC systems
  • Cost estimation for recommended upgrades
  • Batch assessment capability for portfolio-level analysis (useful for housing authorities or large landlords)
Solution Requirements

The MVP should demonstrate the following capabilities:

  • Data capture: An intuitive way to collect building information. This could involve image recognition (photographs of buildings, rooms, and heating systems), a structured questionnaire, integration with publicly available data sources (property records, satellite imagery, and mapping data), or a combination. We are not prescribing the approach. We want to see what innovators come up with.
  • AI-powered assessment: Use of AI/ML to process the captured data and produce an energy rating estimate. The system should be transparent about its confidence level and flag where assumptions have been made versus where hard data was available.
  • Rating output: A clear, understandable output showing the estimated BER grade (A–G), the energy performance indicator (kWh/m²/yr), and the key factors driving the rating.
  • Improvement recommendations: Suggestions for upgrades that would improve the rating, similar to the Advisory Report that accompanies a standard BER certificate.
  • User experience: The app should be usable by non-technical users, including homeowners, real estate agents, and energy advisors. It should work on mobile devices.

The submission should include:

  • Working MVP: A functional application that demonstrates your approach. This can be a web app, mobile app, or API with a front-end demo. It should work with the sample data provided by EnergyBright and produce energy rating estimates.
  • Technical documentation: A clear explanation of your approach, including what AI/ML techniques you used, how you handle data capture, what assumptions your model makes, and how you validated accuracy against the sample BER data. Keep it concise. We value clarity over volume.
  • Architecture overview: A brief document or diagram showing how your system is structured, with specific attention to how it could be extended to support additional European EPC frameworks.
  • Demo video: A short video (5 minutes maximum) walking through the app from a user's perspective. Show us the experience, not just the code.
  • Team profile: Brief bios of each team member, highlighting relevant experience in AI/ML, building science, software development, or related fields.
Constraints and Limitations

Some practical realities to factor into your design:

  • Incomplete information: In the real world, not all building information will be available. A good app should handle uncertainty gracefully, using sensible defaults where data is missing and being transparent about it.
  • Photography challenges: If your approach uses computer vision, consider that building photos will vary wildly in quality, lighting, angle, and what is actually visible. Insulation, for instance, is usually hidden behind walls.
  • Age-based defaults: Much of BER assessment relies on assumptions tied to construction era. A building from the 1970s has different default insulation values than one from 2010. Your model should account for this.
  • Retrofit documentation: Many buildings have been upgraded (new boilers, added insulation, replacement windows), but documentation may be patchy. The app should provide ways for users to indicate upgrades even if they lack precise specifications.

SDG Alignment
SDG

Target 7.3: Double the global rate of improvement in energy efficiency by 2030. Building energy use accounts for approximately 40% of total energy consumption in the EU. Accelerating BER and EPC assessments directly enables retrofit decisions that reduce this figure. By making energy assessments faster and more accessible, this challenge removes a key bottleneck to household-level energy efficiency improvements at the scale required to meet the 2030 target.

SDG

Target 11.6: Reduce the adverse per-capita environmental impact of cities, including air quality and waste management. Poorly insulated homes are a leading contributor to urban CO₂ emissions and fuel poverty. Rapid energy assessments enable targeted retrofit prioritization, allowing local authorities and housing agencies to identify and upgrade the highest-impact buildings first. This challenge supports the systemic urban transition needed to meet SDG 11.6 in European cities.

SDG

Target 13.2: Integrate climate change measures into national policies, strategies, and planning. EU countries now embed building energy performance rules in national climate laws, and the 2024 EPBD revision sets binding retrofit timelines that depend on large-scale EPC assessments. This challenge helps turn those commitments into real emissions cuts by removing the assessment bottleneck that currently blocks implementation.

Challenges.one Logo

Multiverz © 2026