March 28, 2023

How to onboard Chat GPT into your development workflow

I wanted to share some thoughts on Chat GPT and generative AI because here at Reason we’ve undergone something of a radical step change in how we work recently because of it.

I wanted to share some thoughts on Chat GPT and generative AI because here at Reason we’ve undergone something of a radical step change in how we work recently because of it. In fact, I can say it has changed my workflow entirely! More than that: I’m producing more and the quality is better!

How can a bot have that much of a profound impact on the work of a senior software engineer with so many (more than I want to say) years experience behind him you ask?

Well it's not just me - everyone I know and speak to about it is finding the same thing. In fact, if you aren’t embedding it into your workflows now, you're going to get left behind because those that are will steal a march on you in terms of productivity and output.

In terms of advantages we’re finding a significant acceleration of projects because we can start much later in the process because the AI has done the ground work. It’s only left for us to review and refine and adapt and iterate.

It also means the team is free to focus on more innovative tasks within the project with a lot of the basic coding done in advance.

But in terms of the lessons we’ve already learned and the principles we’re applying in order to approach a more systemic use of Chat GPT in our work, I thought it would be useful to share these principles for those not yet used to embedding it within their processes:

  1. Identify use cases: Determine the specific tasks where ChatGPT can provide the most value, such as debugging, designing, writing boilerplate code, writing business logic, code review, testing, and documentation.
  2. Team training: Provide training sessions to familiarize the team with ChatGPT's features, capabilities, and limitations. This will help them understand how to use the tool effectively and get the most out of it.
  3. Establish guidelines: Create guidelines on when and how to use ChatGPT. Set expectations for how the tool should be used and its role in the development process, e.g: what kind of questions it can answer, and how to escalate issues that cannot be resolved by the chatbot.
  4. Integration with development tools: Integrate ChatGPT with your existing development tools, such as IDEs, code repositories, and project management software (such as Jira or Trello) to streamline its usage within the team's workflow and receive notifications when new tasks or issues are assigned.
  5. Set up monitoring and feedback loops: Monitor the team's usage of ChatGPT and gather feedback to identify areas of improvement or additional training needs. Regularly review the tool's performance to ensure it continues to meet the team's requirements.
  6. Encourage collaboration: Promote a culture of collaboration, where team members can share their experiences, insights, and tips for using ChatGPT effectively. This can help the entire team learn from one another and make better use of the tool.
  7. Regular updates: Stay informed about updates to ChatGPT and share them with the team. New features, improvements, or bug fixes can enhance the tool's usefulness and address any limitations.
  8. Evaluate impact: Periodically evaluate the impact of ChatGPT on the team's productivity, code quality, and overall development process. Use this information to make adjustments or expand the tool's use if it is providing value.

Btw, Chat GPT wrote this guide, but I entirely agree ;)

To get a sense of what’s next, watch this to see what’s in store with GPT4.

I’d love to hear how you're using Chat GPT in your workflows, and if you want to talk in more detail about how it can help you in specific projects, please get in touch!