Prompt Engineering for Performance and Sustainability

How you prompt AI matters. It directly impacts output quality, speed, accuracy, and even sustainability. In a recent webinar, Scope3 CEO Brian O’Kelley shared practical techniques for prompt engineering and model calibration, followed by a hands-on walkthrough of Scope3’s agent builder.

Missed it? Here are the top takeaways, or watch the full recording here.

1. Not all prompts are created equal

There are four basic types of prompts:

• Prescriptive: A specific command (“Do X, then Y, then Z”). Helpful when you know exactly what you want.
• Descriptive: A general goal (“Give me a competitive analysis”). These prompts give the model room to interpret and infer.
• Imaginal: Asking the AI to embody a role (“You are a seasoned market researcher…”) to guide its reasoning.
• Illustrative: Providing examples of good and bad outputs to steer the model more precisely.

In a live demo, the same task — analyzing cola brands — produced wildly different results depending on the prompt structure. The most effective output came from a prompt that combined imaginal framing and illustrative guardrails.

Try this: Experiment with the same prompt across all four styles. Then refine based on output quality, speed, and alignment with your goal.

2. Efficient prompts = lower carbon

Every prompt carries a resource cost: energy, water, and carbon emissions. Scope3’s AI Measurement Platform quantifies these impacts by analyzing the size of your input, the model being used, and the location of the data center that processes your prompt.

One surprising insight: Removing “please” from your prompt reduced 7 milliliters of water waste and shortened the number of tokens (i.e., words or word-parts) being processed.

Tip: Prompt efficiently. Add length and complexity only when it improves the output. Shorter, clearer prompts save time and emissions.

3. Calibration is the secret to building reliable brand agents

In the second half of the session, O’Kelley demonstrated how to build a brand suitability agent using the Scope3 Agentic Media Platform.

Step by step, he showed how to:

• Prompt the model to write brand standards (e.g., Nike’s values and tone)
• Apply those standards to content
• Analyze outputs with AI-generated “reasoning”
• Refine the prompt and calibrate performance using human feedback

A good prompt is only the starting point. Ongoing calibration ensures the model stays aligned as content types evolve or brand needs change.

Prompt + test + tune = better outcomes. Use structured calibration to train the model on what’s aligned (or not) with your brand.

4. Sustainability and performance aren’t at odds

Efficient AI is better for the planet and for business. When prompts are focused and well-structured, they require fewer inferences, generate faster outputs, and deliver results that need less human intervention. The same choices that reduce emissions also improve clarity, speed, and quality.

Effective prompting minimizes waste in both energy and effort. By using only what you need — and refining as you go — you unlock the full value of AI without introducing unnecessary complexity or cost.

The best prompt is the one that gets you to the right answer quickly and sustainably.

Want to build your own agent?

Brand Standards is live in the Scope3 platform. As you build your agent, remember: optimizing for performance, safety, suitability, or sustainability starts with a smarter prompt.