With the popularity of chatGPT exploding to millions of users in a matter of few days, one question that has grabbed everyone’s attention is how chatGPT, or similar AI tools impact our daily lives, especially at work. Well, for starters, such tools are already being used to make work easier for many roles, from developers to writers. Product managers have marched ahead with chatGPT helping them come up with elaborate & engaging release notes.
But the question is – is this AI assistance needed at all or is it just a flash in the pan?
The need for assistance in creating release notes
One of the fundamental things that release notes do is to articulate the impact of the product changes on the stakeholder experience. Stakeholders could be end-users, internal support or marketing team & more. Release notes an essential marketing tool that keeps customers abreast of the progress made on the product roadmap. Compiling release notes is a detail-oriented and repetitive task and may involve analyzing & aggregating data from multiple different sources. Often, product managers are tasked with writing release notes. While their contextual knowledge and domain expertise helps them compile these release notes, their time can be put to better use in other activities of their core competence. The amount of effort involved in writing release notes after every product update is often miscalculated; from adhering to an existing template (or creating a new one if none exists) to identifying highlights for different target audiences, to translating them into multiple languages (for products used internationally), the variations are many. Even if each variation takes just 20 minutes, creating 5 of them means nearly two hours spent on a task that can be mostly automated.
How AI tools can help
With the large language models based AI tools, it is easier than ever to ‘generate’ content. How relevant & useful the AI generated content is, that depends on the kind of input or ‘prompt’ you provide to the AI.
Few ways in which generative AI tools like ChatGPT or Bard can help product managers in release notes generation:
- Revising the release notes to match the brand voice & tone
- Automatically translating release notes into multiple languages
- Refining the changelog content to align it with the target audience & distribution channel/s
- Coming up with the first draft of release notes to prevent writer’s block
Using ChatGPT to write release notes
The effectiveness of ChatGPT (and other AI tools) depends on prompts fed by the users. More detailed & specific prompts will result in a better output from the AI tool. In your prompts, make sure to include details like name of your product, version identifiers, release date and the most important one, the list of changes being introduced as part of this release.
Here are a few ways to make release notes better with ChatGPT:
Understand the tone
In which release notes should be written. Organizations with established persona in the market need to talk a certain way, and giving details about this in the prompt is essential. Some probable options for tone are – Casual, Friendly, Professional, Witty & so on. One can even provide sample release notes in the prompt & then ask ChatGPT to learn from the sample data.
Focus on the user’s needs
while writing prompts for release notes. Comments on user persona, proficiency level, or the feature’s complexity can make the AI-generated release note more in line with what users expect instead of being a generic list of changes and feature benefits.
Start with a template
even though ChatGPT can create one based on the inputs provided. Having a structure helps the tools to group information better, and release notes owners have to do less formatting if the organization uses the template regularly. If you are already using Automated release notes for Jira, we have got you covered here. Just getting started you’re your release notes/changelog processes? No problem. Here are some beautiful release notes examples to inspire you.
Review and revise the output
Before publishing to ensure that they are not too generic and are consistent with your expectations. ChatGPT can rewrite or rephrase the input (release notes, in this case) in a specific tone, be it a leading world figure, the organization’s CEO – or any fictional character, but it can get verbose in places it doesn’t need to. A glance at the content is enough to identify the areas for trimming.
Use Google Translate and ChatGPT in tandem
to improve the accuracy of releases notes translations. The advances in NLP have furthered the translation capabilities of the likes of Google Translate. ChatGPT can also understand the context in which the document needs to be translated. But there is still a chance that the message gets lost in translation. Thus, it is wise to compare outputs from both Google translate & ChatGPT to collect best of both the worlds. Whether you use Google translate or something else, as long as you are putting the effort to improve – you will hit the mark. ChatGPT is capable of translating to multiple languages (the count stands at 95, as of now), product teams that serve multiple geographies can extract maximum value while authoring release notes – and send them along to be verified by translators who understand the nuances of the language.
Conclusion
ChatGPT can greatly help product development teams by simplifying their technical writing needs, mainly release notes. Not only can team members focus on their core expertise areas, but they also get value added assistance from the AI. While it can be easy and practical to use the tool as a source of information and assistance, ChatGPT is still in its infancy and has plenty of drawbacks and limitations. Starting with information accuracy, there are a lot of essential aspects that need to be addressed, privacy being a prominent one among them. Teams must exercise caution while using the tool – it is better to stick with routine tasks like writing release notes. Even then, the output from ChatGPT should be used as the base on which the release notes are built – so that teams can ensure user needs are addressed.
