Get more from AI without spending more

AI usage gets expensive fast. Better prompting, reusable frameworks, prompt logs, and guardrails help teams get more, and spend less.

By Steve Petersen,
Marketing Technologist, Wyndham Hotels & Resorts

Published on 2026-10-01T00:00:00Z

Share this article

Table of Contents

Many enterprises are seeing their employees rack up large AI bills, especially through the use of agents in platforms. Sometimes, at a cost significantly greater than the value of the productivity gains. However, employee AI usage doesn’t have to cost more than it returns.

Martech users are no exception, as they try to improve processes. However, AI tools alone can’t do that.

There are several ways to help ensure that AI tools are a wise investment for organizations. These tactics can help individuals and organizations get more value from AI without wasting time or money.

How to make AI prompting more efficient

Verify AI output before using it

It is so critical to verify AI output that most LLMs advise users to double-check it to ensure it makes sense.

For example, I recently used Claude to develop “Would You Rather” questions for a casual work team hangout. One gem was: “Would you rather find $10,000 but have to give half to a stranger, or find $1,000 and keep it all?” Hint: Do the math. Why not keep $5,000 and do a nice thing for someone else instead of simply netting $1,000?

This may seem obvious, but it can never be said too often.

Set guardrails for AI use

Employees who vibe code need clear guidelines to protect the organization from costly mistakes and liabilities.

Scott Brinker, who coined the term martech, is a big booster for marketers using low- and no-code tools and practices like vibe coding to accelerate their work. However, he also knows that guardrails are necessary, noting, “Well-governed vibe coding is not an oxymoron.”

Be strategic about where you prompt

Prompting comes at a cost, so you need to know where to do it cost-effectively. While many platforms offer prompting features, it’s essential to consider the cost of credits. It may make more sense to refine a prompt on another platform, such as Google Gemini, ChatGPT, or Claude, before using it.

Iterate where it’s cheaper before using the more expensive option.

Use prompt frameworks to reduce waste

There are numerous prompt frameworks that can help users form a more comprehensive perspective from the beginning. Here are just a few of them worth considering for marketing contexts: COAST, CO-STAR, FOCUS, and MARK.

  • COAST (context-optimized, audience-specific tailoring) uses well-defined prompts to generate outputs tailored for distinct audiences and situations.
  • CO-STAR (context, objective, style, tone, audience, and response) is good for marketing journeys. Accounting for these six aspects of any journey will help refine a journey’s definition before any AI credits are spent.
  • FOCUS (function, outcome, context, usage, and specific) is meant to build prompts that will yield outputs suitable for practical applications. For example, this works well for prompts meant to generate customer-facing content.
  • MARK (market, audience, research, and key actions) is ideal during the strategic phase of marketing planning — like selecting channels.

Prompting frameworks help define the intent behind the exercise. When the speed of production outpaces the requisite human deliberation, use frameworks that define requirements before spending AI credits.

Build a library of effective prompts

Prompt logs are where individuals record the prompts they’ve used. They should include the quality of the resulting output to identify prompt patterns to avoid or mimic. Here are some of the other benefits:

  • First, by tracking the development process, they can aid audits and internal controls. Many critical audits may soon require them.
  • Second, they mean less “reinventing the wheel.” A successful prompt can serve as a template for others. It doesn’t make sense to start from scratch for every vibe coding endeavor. This can help save time and AI credits.
  • Third, they facilitate team development. New team members can use the prompt library to learn what works. Thus, they can build upon past successes. This can help the team function and perform well.
  • Fourth, the logs are great deliverables for vendors and contractors. In addition to providing documentation on how to maintain what was produced, they can help determine whether it meets contractual requirements for payment.

Vibe coders should also refer to these logs.

When employees are vibe coding, have them check the prompt log before starting from scratch. There may already be a prompt they can adapt with a few small changes.

They should also document the prompts that work well. As people get better at prompting, they’ll develop approaches and templates that others can reuse.

Use vendors’ AI enablement resources

Ask vendors for resources regarding effective prompting in their platforms.

Tech vendors have a strong incentive to help their customers use agentic features effectively. This is especially true after the recent SaaSpocalypse scare, which demonstrated how AI can make the build side of the buy-versus-build debate more viable.

Providing robust agentic features and enabling customers to use them is a critical tactic for preventing churn. They don’t want customers to build their own solutions or move to a competitor that makes that easier.

Ask your account teams for the enablement resources they have. That will help your organization write prompts more efficiently and effectively.

Keep building your AI skills

Workplace applications of AI are constantly changing. Find ways to develop the new skills needed to master them.

Organizations should train their employees, while individuals are accountable for developing their professional skills — regardless of whether their employer pays for training.

There are numerous inexpensive professional development options for individuals to explore. These include asynchronous resources, such as books, websites, videos, and podcasts, as well as in-person and virtual meetups, vendor training, and fellow professionals.

Completing an inexpensive prompt-engineering course, for instance, will definitely help martech practitioners deliver more value to their employers. Don’t forget to take advantage of employer-provided offerings.

Make AI efficiency an ongoing practice

Prompting is relatively new to most of us. So much is changing as this new professional discipline develops. It’ll continue to evolve as all disciplines do.

While new tactics will emerge, these practices should help individuals — like martech practitioners — and organizations use AI more efficiently and effectively and get more value from the time and credits they invest.