AI Sustainability Policy

Last updated: October 16, 2026

AI can do a great deal of good, but every model run uses electricity, hardware and water for cooling, and that has a cost for the environment. Cyberdune Technologies Pvt. Ltd. ("Cyberdunes", "we", "us" or "our") believes a company that builds with AI is responsible for using it carefully. This AI Sustainability Policy sets out how we use AI efficiently, how we manage the carbon footprint of our products and operations, and the standards we hold ourselves to.

It applies to Apex^®, our AI-native digital experience platform (DXP), the services and Masterclass programs we deliver, the website at www.cyberdunes.com, and how we run our business.

Our Principles

• Use AI where it adds real value. We use AI to solve a clear problem for a customer, not because it is available.
• Do more with less compute. The most sustainable computation is the one that never has to run.
• Measure before we claim. We base environmental statements on data, and we do not make claims we cannot support.
• Keep people in charge. Responsible AI is sustainable AI: accurate, reviewed and fit for purpose, so work is not wasted or repeated.
• Think about future generations. We weigh the long-term environmental cost of what we build, not only the short-term benefit.

How We Use AI Efficiently

We design Apex^® and our internal workflows to keep the energy used by AI as low as practical:

• Right-sized models. We match each task to the smallest model that does it well, and reserve the largest models for work that genuinely needs them.
• Reuse instead of regenerate. We cache and reuse results, such as research, summaries and structured data, rather than asking a model to produce the same answer again.
• Batch and schedule. Non-urgent jobs, such as monitoring how AI engines describe a brand, run in batches on a sensible schedule instead of continuously.
• Focused prompts and context. We send models only the context a task needs, which reduces the computation behind every request.
• Quality checks before publishing. Built-in reviews catch errors early, so content does not need to be regenerated again and again.
• No AI for its own sake. Where a simple rule, template or search does the job, we use it instead of a model.

Lean, Low-Carbon Websites

Every page that is downloaded uses energy in data centers, networks and devices. We build websites, including our own, to be light and fast:

• pages are pre-built and served from a global network close to visitors, instead of being generated on every visit;
• images are compressed and sized for each screen, so phones do not download images meant for large monitors;
• fonts are self-hosted and only the characters a page needs are loaded;
• we avoid heavy tracking scripts, advertising tags and unnecessary third-party code; and
• we measure page weight and Core Web Vitals and treat regressions as defects.

Faster, lighter pages are better for visitors and for search, and they use less energy every time they load.

Our Carbon Footprint

Most of our footprint comes from the cloud computing and AI services we use, the devices our team works on, and travel to deliver training and client work. We are committed to:

• Measuring the main sources of our emissions, including cloud and AI computing, using recognized greenhouse gas accounting methods and the data our providers make available;
• Reducing them first, through the efficiency practices in this policy, before relying on offsets;
• Choosing responsible providers, preferring hosting and AI providers that publish their environmental performance and are working toward renewable energy; and
• Reporting honestly, sharing progress once we have reliable data, and never describing our products or company as carbon neutral without independent evidence.

Responsible Operations

• Hardware: we extend the life of laptops and devices, and recycle electronic waste through authorized e-waste recyclers in line with India's E-Waste (Management) Rules.
• Travel: we run meetings and many training sessions online where that works as well, and plan in-person work to avoid unnecessary trips.
• Paperless by default: proposals, invoices and certificates of completion are issued digitally.
• Office habits: we switch off idle equipment and favor energy-efficient devices.

Helping Our Customers Do the Same

Apex^® helps marketing teams publish clear, well-structured content that search engines and AI assistants can understand the first time. Clear entities, structured data, llms.txt files and XML sitemaps with IndexNow help crawlers find changes without repeatedly fetching every page, which reduces wasted crawling. We also share our efficient practices with clients and in our Masterclass programs, so the teams we work with can build responsibly too.

Responsible AI Commitments

Sustainability and responsibility go together. We keep people in control of AI-assisted work, review content before it is published, respect copyright and privacy, and do not use customer data to train public AI models without permission. Our Privacy Policy explains how we handle personal data.

Governance and Review

Our leadership team owns this policy. We review it at least once a year, and whenever we make a significant change to how we use AI or run our infrastructure. We welcome questions, ideas and challenges from customers, partners and our team.

Contact Us

To ask about this policy or share an idea, email support@cyberdunes.com or write to Cyberdune Technologies Pvt. Ltd., Mumbai, Maharashtra, India.