How to Effectively Apply AI in Your Work in 2026
In the rapidly evolving landscape of 2026, many professionals struggle to leverage AI effectively. This article provides essential, counter-intuitive strategies. From using AI to evaluate your ideas rather than generate them, to mastering multi-LLM prompting, and even leaning into AI "hallucinations".
Begin with your own ideas - don't rely on AI to generate ideas for you; instead, use AI to assess and evaluate your ideas
Generative AI is a highly useful tool, but it is not capable of producing truly original ideas. Its context, processes, and learning are based on analyzing and synthesizing existing ideas from others. Due to this inherent limitation, it is important to avoid relying on AI to generate ideas for you. What it produces will largely be variations of existing concepts, rather than truly differentiating insights.
In 2026, consider having generative AI tools evaluate your ideas by providing information on current market conditions, including deals, reasoning, and scope. Then, prompt them to review your ideas in the context of market trends, product capabilities, and potential positioning.
Let AI talk the lingo - Utilize multiple LLMs to assist and improve prompting
One of the key advantages of GenAI is its ability to engage in natural language conversations. However, when delving into complex details, a gap can sometimes emerge between the prompt’s intent and its execution. This is often where additional prompt engineering expertise is needed to guide the process toward the desired final outcome.
Instead, consider utilizing large language models to better understand and craft prompts for other LLMs. Develop a thorough prompt to the best of your ability, then instruct the GenAI to generate a prompt for another GenAI to follow. This approach encourages clear communication with other LLMs through effective prompt engineering, without requiring you to be an expert in the field.
Don’t trust the output - Some AIs will fake it to make something appear correct
Some impressive tools like Replit seem almost like magic. You input your idea for an app, and it generates a solution that appears to make it happen. At least, it seems that way. Currently, technology achieves about 60% accuracy in producing code. However, many large language models (LLMs) have been found to create extensive self-prompts instead of actual code to facilitate the process. As a result, instead of receiving standalone code that can be published directly, you'll often need to publish and utilize LLM credits to make the website functional.
It is important to determine the final technology stack and intended deployment method before starting to create with these tools. Understanding the limitations of the target stack and the specific code requirements upfront can help avoid significant rework later. Additionally, consider whether you want to be committed to hosting on a particular platform, as many functionalities are dependent on that platform’s AI infrastructure.
“Tell Me Why" (Backstreet Boys reference) - Lean into the hallucinations, ask the LLM why it got to that answer
Oh hallucinations. Hallucinations are a common occurrence, often when an AI generates its own role, presents statistics, or attempts to showcase more information than is actually available. Despite careful training, large language models can still produce occasional inconsistencies or inaccuracies. It sometimes feels as though the model has gone off on a tangent or is trying to present itself better.
Instead of dismissing these hallucinations, consider asking the AI why it arrived at that conclusion. You may find valuable insights by requesting the model to explain its reasoning. In my experience, some interesting explanations relate to proximity to another context within a vector database or a different meaning from another language that I hadn’t considered. Asking the AI to clarify can help you understand what’s happening behind the scenes.
Empower your creatives - Utilize AI to assist while leveraging human creativity for design, development, and strategy
The most concerning trend in business has been companies attempting to replace human workers with AI. This approach overlooks the fact that AI can primarily replicate existing human tasks. Large companies have begun to reconsider their layoffs, as seen with Salesforce reducing layoffs in late 2025. Since AI has inherent limitations and there is a general lack of public understanding regarding its effective use, there exists a significant gap between users who leverage it well and those who do not.
Now is an excellent time to invest in high-quality creative talent and enable them to enhance their work with AI. Generative AI is a valuable tool for augmenting human capabilities and increasing productivity. As demonstrated in other major technology shifts, companies that focus on investing in their people rather than replacing them are most likely to succeed.
Meet the Author!
Kerrigan Baron (They/Them) - CEO & Founder @ Fidget Labs
Short Bio - Kerrigan Baron is an accomplished technology leader with over 20 years of experience revolutionizing enterprise digital experiences through cutting-edge MACH (Microservices, API-first, Cloud-native, Headless), complex composable architecture, and AI-Enabled Solutions. As CEO & Founder at Fidget Labs, they founded the consultancy with the express goal of gathering elite professionals across finance, technology, and research to deliver strategic solutions that drive meaningful change for clients across the globe with a human-first mindset.