From Documentation Pain to Available Generative AI Solutions: Getting Started

Generative AI (GenAI) tools are increasingly prevalent in the workplace. Gartner estimates that over 80% of organizations will have used or deployed generative AI apps by 2026.

Documentarians, often tasked with vast amounts of technical content, face growing pressure to use recommended artificial intelligence (AI) tools to be more efficient.

While some may cringe at using AI tools and changing up their operations, these apps provide time-saving opportunities—especially among busy small teams and solo writers.

But a crucial question remains: how do you begin?

This blog guides you on identifying a viable use case, assessing your available AI tools, matching the solution to the problem, and taking the next steps to integrate it in your work.

By the end of the post, you will become more confident about using generative AI to solve your pain and clear about how.

Identify and Evaluate Pain Points

A successful partnership with any AI solution requires a valid business reason why. I started using generative to write better and more efficiently.

Like other documentarians, I needed to break down this larger goal into smaller use cases. To do so, I evaluated my work. I considered:

• Content inventory: Assess the available content and its formats.
• Goals: Define the documentation goals. Assess how well they are being met.
• Audience: Identify the readers and clients.
• Effectiveness: Evaluate how well the content meets user needs (e.g. support ticket resolution rates or data quality metrics).
• Processes: Review the current documentation processes. How efficient are they?
• Resources: Analyze the resources available for the documentation efforts. Be sure to consider roles and responsibilities.
• Metrics and feedback: Determine what inputs and feedback is needed to understand the documentation strengths and weaknesses.

Once I got this information, I thought about areas to improve, like:

• Generate new writings.
• Change roles and responsibilities.
• Correct process inefficiencies.
• Increase resources.
• Improve quality based on feedback.

Through this evaluation, I thought about all sorts of use cases:

• Write the next section, “Understand Generative AI for Documentation” to save time constructing it.
• Generate an invoice for a specific client.
• Align this blog’s contents with KnowledgeOwl’s content strategy, particularly, “Writing doc posts.”
• Make sure my content does not exceed 1,200 words to meet a requirement.
• Have an AI tool suggest three different ways to reword this bulleted list for clarity.

With these use cases in mind, it's crucial to understand how generative AI can address these documentation challenges.

Understand Generative AI in Documentation

As documentarians, we need to know how GenAI can impact our work. GenAI, trained on large volumes of unorganized data, uses natural language processing (NLP) to create human-like text. It can automatically write coherent content, summarize source material, and provide a sounding board to consider and refine writing ideas.

However, GenAI comes with drawbacks. It can generate, “hallucinations” due to a lack of expertise or complete instructions. For example, when I asked it to write the “Understand Generative AI for Documentation” section, it omitted the part of being a sounding board.

Additionally, GenAI tools have a persistent memory. So, we need to be careful not to share sensitive information, like clients’ names and prices. 

I eliminated problematic use cases based on my understanding of GenAI. These included writing a section, generating an invoice, and making sure my content was 1,200 words or less. Then, I narrowed in on documentation use cases that better align with the tools I have on hand.

Match the Available GenAI Tool with the Use Case

While it helps to know how GenAI applications function, not all tools are the same. They vary in their training, expertise, and interface. Moreover, not all documentation use cases work best with the first available tool.

I initially tried Microsoft Copilot, a conversational AI designed to assist with various tasks, through the Microsoft Edge browser.

I asked it three different ways to reword my bulleted use case list for clarity. I prompted the AI tool, scrolled through the text, and selected the best option to reword this section.

I found this output cumbersome to digest, so I switched to another available tool, Lex Pro. Lex is a GenAI word processing app, like Google Docs, with expertise in writing feedback.

As a documentarian, I find the Lex Pro app more familiar and easier to use. It would work well for rewording bullet points and aligning this blog content with KnowledgeOwl’s content strategy. Aligning this blog content with the strategy is of higher value to my readers and client than clarifying only a bulleted list. 

Now that I have a use case, I can hone in on the specific solution steps. I am closer to solving a document pain!

Conclusion: Solve the Documentation Pain

Once you have selected a use case and an AI solution, it is time to take the next steps, depending on your tool access. If you need access, consider pitching and showcasing the solution to your manager to get access.

CIO, Chamorro-Premuzic provides seven strategies, including having patience and sticking with small incremental solutions. Using AI to audit existing content for alignment to a well-stated strategy fits this advice and covers a variety of documentation.

As a solo documentarian with tool access, I can immediately implement the AI-assisted content alignment process in Lex Pro.

In the AI settings for each document written in Lex, I can set the context, link the tool to KnowledgeOwl’s strategy, and ask it for feedback.

Thanks to my AI partner, I am more productive with this piece. Without GenAI, I take hours to switch from the document I am writing, to review a content strategy, and go back to my article to figure out what to edit. With the generative AI tool, this task takes minutes, depending on the feedback.

Even better, I can use the Lex tools to build my prompt and have it on hand for future articles. All I need to do is learn how to provide my AI partner with enough information and how to access this instruction next time.