--- title: "Results from our “Using AI in techcomm” 2025 survey" id: "17972" type: "post" slug: "results-from-our-using-ai-in-techcomm-2025-survey" published_at: "2025-05-22T07:00:49+00:00" modified_at: "2025-05-22T09:02:58+00:00" url: "https://www.cherryleaf.com/2025/05/results-from-our-using-ai-in-techcomm-2025-survey/" markdown_url: "https://www.cherryleaf.com/2025/05/results-from-our-using-ai-in-techcomm-2025-survey.md" excerpt: "In April 2025, Cherryleaf conducted its third annual survey into the use of AI and ChatGPT in technical communications. Below is a summary of our findings. Executive summary This year, we had 103 responses from technical communication professionals. The overall..." taxonomy_category: - "AI" taxonomy_post_tag: - "AI" - "survey" --- In April 2025, Cherryleaf conducted its third annual survey into the use of AI and ChatGPT in technical communications. Below is a summary of our findings. ## Executive summary This year, we had 103 responses from technical communication professionals. The overall findings reveal a field actively engaging with, yet still cautious about, Artificial Intelligence. - Cautious adoption - While familiarity is high, deep, integrated use is still developing. Many are in an experimental phase or use AI for specific, limited tasks. - Self-reliance in learning - The vast majority are teaching themselves how to use AI, indicating a gap in formal training opportunities or a preference for self-directed learning in this fast-evolving space. - Productivity v. quality trade-off - AI is seen as a tool for efficiency (drafting, summarising, rephrasing), but concerns about accuracy, hallucinations, and the need for significant human editing are prevalent. - Ethical and security concerns are significant - Worries about copyright, data privacy, and the “black box” nature of some AI models are common. - Job security is a top concern - The potential for AI to devalue or replace technical communicators is a recurring theme, often linked to management’s perception of AI capabilities. - Desire for guidance - There’s a clear demand for best practices, strategies for effective AI use, and understanding how to optimise content for AI interaction. - Organisational lag - Many organisations are still formulating their AI strategies, including whether and how to best implement AI systems that use their documentation. - The “Human-in-the-Loop” is crucial - Most users recognise that AI is currently an assistant, not a replacement, and that human expertise, critical thinking, and editing are essential. - Autonomous AI – The next frontier - Familiarity with autonomous AI agents is much lower, suggesting this is an area for future exploration and development in the field. ## About the respondents The respondents this year were very similar to previous years. This means we can examine the trends over time. “Technical Writer”, and variations of this job title, were the main people who completed the survey. There were also a few documentation managers, students and freelancers. Most work in the software sector (including Cloud, SaaS, mainframe, development, security, accounting, market research, etc.) Country of residence (Top 5): 1. USA 32% 2. UK 17% 3. Canada 10% 4. India 8% 5. Australia 6% Other countries included: Germany, Poland, Israel, Switzerland, Greece, Belgium, New Zealand, France, Denmark, Romania, Norway, Hungary, Netherlands, and Ukraine. Organisation size: - Small business (including “It’s tiny”): 33% - Medium sized: 26% - Large: 23% - Extra large: 17% ### Familiarity with AI We asked how familiar are you with AI and ChatGPT and how it can be used in technical communication? Most respondents have moderate familiarity with AI, but few consider themselves experts. There has been a significant change between 2025 and 2024, as more people become familiar with AI. | Question | 2025 | 2024 | 2023 | | --- | --- | --- | --- | | I know a lot | 22 | 12 | 3 | | I know a reasonable amount | 49 | 39 | 15 | | I know a little, but not enough | 28 | 46 | 73 | | I know nothing | 1 | 1 | 10 | We also asked, have you had any training in how to use AI? - 75% learned AI tools independently - 25% received formal training This is a slight change from 2024, where 79% learned AI tools independently and 21% received formal training. This indicates a gap in workplace upskilling programmes, and a need for more structured training, such as the AI training course that Cherryleaf offers. ### AI use in technical communication We asked, do you use AI (such as ChatGPT) in your role as a technical communicator? 55% use AI on a regular or semi-regular basis, while 15% avoid it entirely, often due to ethical concerns or employer restrictions. Compared to the survey responses from 2024, there has been a noticeable drop in the number of people not using AI at all, and in the number of people using it a lot. | Question | 2025 | 2024 | | --- | --- | --- | | Not at all | 15 | 28 | | I have done some experimenting | 17 | 20 | | I use it occasionally | 16 | 17 | | I use it sometimes | 25 | 20 | | I use it a lot | 30 | 16 | ### How AI is being used by technical communicators Respondents who use AI employ it for a diverse range of tasks: - Content generation and drafting - Creating first drafts of topics, guides, release notes, blog posts - Generating outlines - Scaffolding content - Getting started on a “blank page” - Editing and rephrasing - Rewording complex sentences, UI text - Improving clarity, tone, consistency - Summarising long texts (Slack threads, meeting notes, articles) - Checking for style guide adherence (though some note hallucinations here) - Simplifying verbose English, aiming for Simplified Technical English - Grammar and spell checking - Research and information gathering - Researching new technologies, features, or complex topics - Asking questions about code or specifications - Summarising technical content for quick learning (AWS docs) - As an “interactive Wikipedia” - Coding and technical tasks - Generating code examples (Python, shell scripts). - Building regex expressions - Assisting with HTML, CSS, JavaScript (for Docusaurus sites, customising authoring tool output) - Formatting (Markdown tables) - Automating documentation tasks (scripting in PowerShell) - Process and workflow improvement - Analysing specifications to create conceptual topics - Brainstorming and generating ideas - Comparing content to templates and reorganising - Tagging articles - Creating UI naming conventions - Generating personas - Translation (as a first draft) Notable quotes: - “I use AI to rephrase complex sentences when my brain is stuck.” - “AI helps me generate release notes and check style compliance.” ### Developing chatbots We asked, does your organisation have an AI system, such as a chatbot, that uses information from your user documentation? Compared to the survey responses from 2024, there has been a noticeable increase in the number of organisations creating chatbots: | Question | 2025 | 2024 | | --- | --- | --- | | Yes, and it can be used by anyone | 23 | 9 | | Yes, but it can only be used by our customers | 11 | 8 | | No, but we have one in development | 17 | 16 | | No, but we plan to | 22 | 34 | | No, and we don’t plan to have one | 28 | 32 | ### Using agentic AI We asked, how familiar are you with autonomous AI agents (which can act/”think” autonomously and automate tasks) and how they can be used in technical communication? This is an emerging field, and unsurprisingly not many respondents were familiar with the technology. ### Key questions, concerns, and perceived challenges Respondents expressed a wide range of questions and concerns regarding AI in technical communication: A. Job security and role evolution - Will AI lead to redundancy or take writers’ jobs? (This was a frequent concern) - How will the technical communicator’s role evolve to work effectively with AI? - How can AI help improve work rather than lead to job losses? - Fear of being “replaced by a tool” B. Quality, accuracy, and ethics - Hallucinations and accuracy – How to deal with AI making things up or providing incorrect information - Ethical concerns - Plagiarism and copyright (models trained on potentially stolen content) - Data privacy and security (especially with proprietary information) - Environmental impact (energy and water consumption of data centres) - Transparency – should customers be informed if AI is used in content creation? - Quality of output – AI-generated content often requires extensive editing; its “chatty tone” can be inappropriate - Need for human oversight and critical evaluation of AI output C. Implementation and best practices - How to effectively integrate AI into existing workflows and tools - What are the best practices for using AI in technical communication? - How to write effective prompts - How to train AI on specific style guides or large content sets (RAG, fine-tuning) - How to make documentation “AI-ready” for consumption by AI systems - What specific tools can enhance productivity? D. Skill development and training - How will up-and-coming writers develop skills if they rely too heavily on AI? - What training is needed to stay relevant? - How to evaluate AI-generated content effectively. E. Organisational and strategic concerns - Lack of AI policies in organisations - Convincing management/executives that AI augments rather than replaces human writers - Security of proprietary content when using external AI tools - How to keep up with the rapid pace of AI development F. Specific technical questions - Using AI for IA and search experience - Automating tasks like drafting short descriptions, making message libraries consistent, rewriting legacy text - Connecting disparate tools (Confluence, Flare, code repositories) - Using AI as a linter for automated style guide verification - Automating documentation drafts based on application interface analysis - AI’s ability to find instances needing change in large repositories based on updates ### Potential outlook The survey suggests that technical communication is at a crossroads. AI tools are becoming increasingly integrated into workflows, offering potential for increased efficiency and new capabilities. However, the path forward requires addressing significant concerns about job roles, ethics, and quality. The field will likely see: - A growing need for “AI literacy” among technical communicators. - Development of new skills focused on prompt engineering, AI output evaluation, and managing AI-assisted content workflows. - A shift in focus for some, perhaps towards content strategy for AI, curating knowledge bases for AI consumption, or developing specialised AI tools for documentation. - Increased demand for ethical guidelines and organisational policies around AI use. - Continued debate and exploration of how to balance AI’s capabilities with the value of human expertise and judgment in technical communication. It shows a profession grappling with transformative technology, with both excitement and apprehension that comes with such change. ### More information [Using Generative AI in technical writing (training course)](https://www.cherryleaf.com/training-courses/using-generative-ai-in-technical-writing-training-course/) ### Share this: - [Share on Facebook (Opens in new window) Facebook](https://www.cherryleaf.com/2025/05/results-from-our-using-ai-in-techcomm-2025-survey/?share=facebook) - [Share on LinkedIn (Opens in new window) LinkedIn](https://www.cherryleaf.com/2025/05/results-from-our-using-ai-in-techcomm-2025-survey/?share=linkedin) - [Share on WhatsApp (Opens in new window) WhatsApp](https://www.cherryleaf.com/2025/05/results-from-our-using-ai-in-techcomm-2025-survey/?share=jetpack-whatsapp) - [Share on Bluesky (Opens in new window) Bluesky](https://www.cherryleaf.com/2025/05/results-from-our-using-ai-in-techcomm-2025-survey/?share=bluesky) - ### *Related* ### Leave a Reply[Cancel reply](/2025/05/results-from-our-using-ai-in-techcomm-2025-survey/#respond) This site uses Akismet to reduce spam. [Learn how your comment data is processed.](https://akismet.com/privacy/) #### Previous post [Podcast 158: The Changing Marketing Landscape](https://www.cherryleaf.com/2025/05/podcast-158-the-changing-marketing-landscape/) #### Next post [Towards a Question and Answer world](https://www.cherryleaf.com/2025/06/towards-a-question-and-answer-world/) #### Browse posts by month Browse posts by month Select Month August 2026 (5) July 2026 (7) June 2026 (14) May 2026 (8) April 2026 (6) March 2026 (14) February 2026 (9) January 2026 (9) December 2025 (7) November 2025 (9) October 2025 (10) September 2025 (7) August 2025 (6) July 2025 (1) June 2025 (3) May 2025 (2) April 2025 (5) March 2025 (3) February 2025 (3) January 2025 (2) November 2024 (8) October 2024 (3) September 2024 (2) August 2024 (3) July 2024 (4) June 2024 (1) May 2024 (2) April 2024 (4) March 2024 (1) February 2024 (2) January 2024 (1) December 2023 (2) November 2023 (9) October 2023 (8) September 2023 (5) August 2023 (3) July 2023 (1) June 2023 (1) May 2023 (2) April 2023 (3) March 2023 (2) January 2023 (2) December 2022 (4) November 2022 (1) October 2022 (1) September 2022 (1) August 2022 (1) July 2022 (1) June 2022 (1) May 2022 (2) April 2022 (1) March 2022 (1) February 2022 (1) January 2022 (2) December 2021 (1) November 2021 (2) October 2021 (1) August 2021 (1) July 2021 (2) June 2021 (3) May 2021 (5) April 2021 (3) March 2021 (3) February 2021 (1) January 2021 (2) December 2020 (4) November 2020 (4) August 2020 (3) February 2020 (3) January 2020 (1) November 2019 (3) October 2019 (1) September 2019 (1) August 2019 (3) July 2019 (2) June 2019 (3) May 2019 (4) April 2019 (1) November 2018 (2) October 2018 (1) July 2018 (2) June 2018 (4) May 2018 (2) April 2018 (3) March 2018 (4) February 2018 (2) January 2018 (3) December 2017 (2) October 2017 (3) September 2017 (1) August 2017 (3) July 2017 (2) June 2017 (7) May 2017 (4) April 2017 (5) March 2017 (1) February 2017 (7) January 2017 (1) December 2016 (3) November 2016 (4) October 2016 (1) September 2016 (5) July 2016 (1) June 2016 (3) May 2016 (1) April 2016 (2) February 2016 (2) January 2016 (5) December 2015 (1) November 2015 (4) October 2015 (2) August 2015 (1) July 2015 (1) June 2015 (2) May 2015 (3) February 2015 (2) January 2015 (6) December 2014 (1) October 2014 (1) September 2014 (4) June 2014 (4) May 2014 (1) April 2014 (2) March 2014 (3) February 2014 (4) January 2014 (4) November 2013 (2) October 2013 (5) September 2013 (3) July 2013 (3) June 2013 (11) May 2013 (4) April 2013 (6) March 2013 (7) February 2013 (4) January 2013 (4) December 2012 (9) November 2012 (23) October 2012 (14) September 2012 (8) August 2012 (4) July 2012 (12) June 2012 (12) May 2012 (10) April 2012 (8) March 2012 (12) February 2012 (3) January 2012 (6) December 2011 (2) October 2011 (3) August 2011 (1) June 2011 (5) May 2011 (2) April 2011 (1) January 2011 (1) December 2010 (1) November 2010 (3) October 2010 (4) September 2010 (8) August 2010 (5) May 2010 (2) April 2010 (3) March 2010 (3) February 2010 (2) January 2010 (3) December 2009 (2) November 2009 (6) September 2009 (1) July 2009 (6) June 2009 (7) April 2009 (8) March 2009 (3) February 2009 (3) January 2009 (3) December 2008 (1) November 2008 (3) October 2008 (3) September 2008 (4) August 2008 (1) July 2008 (3) June 2008 (5) May 2008 (2) April 2008 (2) March 2008 (3) February 2008 (2) January 2008 (1) November 2007 (3) October 2007 (2) September 2007 (1) July 2007 (3) November 2006 (1) October 2006 (2) September 2006 (3) August 2006 (1) April 2006 (1) March 2006 (3) #### Browse posts by category