I think like most technical writers I’ve been on the fence as regards using ChatGPT or Gemini to write tech docs. However, I discovered that by using so called vibe-coding I found a new way to get into coding again. And started to really enjoy it. Once that started, I came back to seeing if I could vibe-write using AI, if that makes sense. The answer is yes and no. I don’t actually write in ChatGPT, rather use it to get started, draft, and refine. Something I have found helpful is to save a list of prompts that I can tweak for different docs. This article touches on how I do this. See if it helps.
The idea of using AI to write technical documentation often triggers understandable reservations. Technical writers, engineers, and product teams are trained to value precision, accountability, and traceability—areas where AI is frequently perceived as weak or inconsistent.
Common concerns include:
- AI introducing subtle inaccuracies that are difficult to detect
- Output sounding generic, voiceless or disconnected from the actual product
- Loss of control over terminology, tone, or scope
- Difficulty passing technical, QA, or compliance review
These concerns are valid. AI systems generate content based on patterns, not understanding. Without constraints, they can produce text that appears confident and coherent while being incomplete, misleading, or factually incorrect.
For this reason, AI should not be treated as a replacement for technical documentation practices.
Instead, it should be used as a drafting aid, operating within a clearly defined structure and under human supervision (aka human-in-the-loop).
Why User Guides Are Well Suited to Controlled AI Use
User guides follow a predictable and widely accepted structure: purpose, scope, system description, procedures, error handling, and reference material. This makes them particularly suitable for AI assistance, provided that structure is established in advance.
AI performs best when asked to generate content for well-defined sections with clear intent.
It performs poorly when asked to decide what belongs in the document or how much detail is sufficient. Those decisions remain the responsibility of the tech author. Templates therefore play a central role in ensuring AI is used effectively and safely.
The Role of AI in Writing User Guides
What AI Does Well
AI tools such as ChatGPT and Google Gemini are effective at:
- Producing first drafts of clearly scoped sections
- Expanding bullet points into readable prose
- Rewriting text for clarity and consistency
- Generating structured lists when given explicit instructions
These tasks are time-consuming for humans but low-risk when the output is reviewed carefully.
Where AI Needs Clear Constraints
AI is less reliable when asked to:
- Define scope or exclusions
- Invent technical or architectural details
- Interpret regulatory, security, or privacy requirements
- Replace subject-matter expertise
For this reason, AI should always be used within an existing document framework rather than as a free-form authoring tool.
Why Templates Matter More When Using AI
Using AI without a template often results in content that is readable but incomplete, inconsistent, or difficult to review. A professional template provides the structural discipline that AI lacks.
A user guide template ensures:
- All required sections are present
- Content is organized consistently
- Reviewers can assess completeness efficiently
- Missing or misaligned information is easier to identify
When AI is used inside a template, the template becomes the controlling mechanism and AI becomes
a drafting assistant rather than a decision-maker.
The Klariti User Guide Prompt Template Pack
The Klariti User Guide template pack reflects established technical documentation practices and includes sections for:
- Purpose and scope
- System description and key features
- Installation, access control, and startup
- Step-by-step instructions
- Errors, messages, and recovery
- Supporting appendices and references
Each section includes guidance that makes it straightforward to write focused AI prompts aligned to the document’s intent.
Recommended Workflow: AI + User Guide Template
To use AI to help draft a user guide, follow this pattern:
- Open the User Guide template and select a specific section
- Use a focused prompt designed for that section
- Paste the draft output into the template
- Review and revise the content critically
- Repeat this process section by section
This approach keeps AI constrained, improves reviewability, and reduces the risk of structural gaps.
Example AI Prompts Mapped to User Guide Sections
The following prompts are taken from the User Guide Prompt Library and mapped directly to sections in the template.
Purpose (Prompt ID 1.1)
Draft a purpose statement for the [Software Name] user guide.
Summarise the system’s function, intended users, and the objective of this document.
This prompt works well because it focuses on intent rather than detail. The output should still be checked to ensure it reflects actual usage rather than aspirational language.
Scope (Prompt ID 1.2)
Draft the scope section for the [Software Name] user guide.
Include what is covered, what is excluded, and any security or privacy considerations.
AI-generated scope statements should always be reviewed carefully, particularly exclusions, which often require organisational judgment.
Key Features (Prompt ID 2.1)
Describe the key features of [Software Name] from an end-user perspective.
Focus on practical capabilities rather than internal architecture.
This prompt is effective for producing user-oriented descriptions, provided feature accuracy is validated.
Step-by-Step Instructions (Prompt ID 4.1)
Create step-by-step instructions for completing [Task Name] in [Software Name].
Assume the user has basic access but no prior experience.
This is one of the strongest use cases for AI, though warnings, prerequisites, and edge cases typically require manual refinement.
Before and After: AI-Assisted Drafting Example
Before (Manual Draft)
The purpose of this document is to explain how users can install and use the reporting module.
It also describes basic troubleshooting steps and access requirements.
After (AI-Assisted Draft, Then Reviewed)
The purpose of this document is to provide end users with instructions for installing, accessing, and using the reporting module of [Software Name]. The guide is intended for operational users and administrators who require a practical reference for day-to-day tasks. Basic troubleshooting and access control considerations are included where relevant.
The AI-generated version provides more structure and clarity. The final wording should still be reviewed to confirm accuracy and alignment with product terminology.
Reviewing and Refining AI Output
AI-generated content should be reviewed using the same standards applied to any draft:
- Verify factual accuracy
- Remove vague or ambiguous language
- Align terminology with product and organisational standards
- Confirm consistency across sections
AI accelerates drafting, but accountability remains with the author.
Encouraging Responsible Experimentation
AI is most effective when used experimentally and incrementally. Rather than attempting to automate an entire user guide at once, start with a single section and a single prompt.
Over time, this approach naturally leads to the creation of a prompt library tailored to specific products, users, and documentation standards. Such a library is far more valuable than any individual AI-generated output.
Next Steps
AI can be a useful assistant for writing user guides, but only when used with structure, restraint, and judgment. Templates provide the structure. Prompts provide direction. Human review provides trust.
Used together, these elements allow documentation teams to work more efficiently without compromising professional standards.
Have you tried to use AI to refine the tech doc process?
Want more AI Writing Tutorials? Sign up for our newsletter to receive more practical AI writing tips and tutorials delivered straight to your inbox.

