top of page
  • Nikita Solodkov

Understanding developers who build generative AI applications

While the idea of generative artificial intelligence (AI) is not new, a lot of progress has been made over the past few years that has allowed software developers to elevate the functionality of their applications by incorporating AI. In recognising that AI will only grow in popularity, vendors offering generative AI services must adapt to the specific needs of this technology to maintain a competitive edge in the industry.


This blog post aims to provide a basis for understanding developers who build generative AI applications by exploring the technologies they use and the information sources they prefer. The analysis is based on data collected from over 16,500 respondents who answered questions about incorporating generative AI functionality into their applications in the 25th edition of our global Developer Nation survey, which was fielded in Q3 2023. More information about how developers create generative AI solutions can be found in the accompanying report.


Cloud computing is essential for many generative AI applications

In recent years, cloud computing has become a strategic necessity for many modern software projects. While hosting software locally has its advantages, running software in the cloud allows developers to build scalable solutions that can accommodate demanding workloads.


Due to the computationally intensive nature of advanced generative AI models, cloud computing has become particularly important for developers whose applications incorporate generative AI functionality (e.g. through Amazon Bedrock, Google Vertex AI, etc.). According to our survey data, 66% of these developers use cloud environments to deploy their code. In contrast, only 50% of those who do not build generative AI applications run their applications on cloud servers, for whom end-user devices are the most popular choice.

data graph on code execution environment

Let us now take a closer look at the developers whose applications run on the cloud. The relative usage rates of public and private clouds are very similar for both groups of developers. However, we observe a significant difference in the adoption of multi-cloud solutions between those who add generative AI functionality in their applications and those who do not (32% vs 18%).


This indicates that developers who build generative AI applications are more willing than their counterparts to use different cloud computing services for different workloads. As such, these developers may be more willing to adopt different cloud providers if they provide the features they desire. Furthermore, these developers may also be more likely to choose generative AI services that do not restrict them to a given vendor’s tooling ecosystem and gravitate towards products that allow easy integration with technologies offered by various vendors.

data graph on cloud environment types developers use

Developers who build generative AI applications are far more likely than average to use low-code / no-code tools


Beyond this, we also see that 75% of developers who build generative AI applications use low-code or no-code (LC/NC) tools in their workflows. Conversely, only 41% of those whose applications do not utilise generative AI functionality report using such tools. This suggests that it may be highly advantageous for vendors offering generative AI tools to offer LC/NC capabilities along with code-based SDKs and APIs.

data graph on adoption of low-code / no-code tools

Breaking down this information by experience reveals that this pattern is extended across all experience levels. In broader research, we find that those with more than ten years of experience are the least likely to LC/NC tools. Despite this, 59% of these developers who also build generative AI applications report using LC/NC tools to some capacity in their projects. This is significantly above that of their counterparts who do not build generative AI applications but have the same experience (29%). Not only that, but it also exceeds the usage of LC/NC tools amongst the less experienced developers who do not engage with generative AI development.


This suggests that LC/NC tools are highly desirable for building generative AI applications, even by the most experienced developers. In turn, this opens up a wide range of opportunities for vendors offering generative AI products to enhance developer experience and gain new customers through offering LC/NC capabilities.

data graph on adoption of low-code / no-code tools by experience

How to reach developers who build generative AI applications

While it is important to understand what developers look for in tooling, capturing their attention is just as crucial in building a competitive advantage in the industry. One way to look at this is by considering where they go to find information and updates about software development.


Our research reveals that developers who build generative AI applications prefer to use community-driven sources of information. For instance, 44% of these developers turn to open source communities and a further 38% to community websites and forums for updates and information about software development. This highlights how important it is for vendors offering generative AI products to maintain a good relationship with the community. In turn, this allows them to effectively reach their target audience through methods that extend beyond the vendor-owned resources.


Comparing this to those whose applications do not utilise generative AI functionality shows that these developers are less likely to rely on community resources overall. Instead, developers who build generative AI applications are significantly more likely to engage with conversational AI services (37%) than their counterparts (25%). This suggests that there is a link between developers using generative AI services and adding generative AI features to their own applications.



data graph on where developers go to find information

In addition to communities and chatbots, social media is very popular amongst developers who build generative AI applications. Social media is a great resource for getting updates about new technologies. In broader research, we observe that the popularity of social media is typically highest amongst developers in the early stages of their careers, decreasing greatly as they gain experience. However, this is not the case for those who build generative AI applications. For these developers, the use of social media for information remains at around 38% across all levels of experience. This suggests that when it comes to generative AI, using social media can be just as effective at reaching experienced developers.


data graph on usage of social media for information by experience

Key takeaways:

  • Cloud computing: 66% of developers who build generative AI applications use cloud environments to run their applications, compared to 50% of other developers.

  • Low-code/no-code tools: There’s a significant trend of generative AI developers (75%) using low-code/no-code tools, indicating a market opportunity for vendors.

  • Information sources: Developers who build generative AI applications prefer community-driven information sources, suggesting vendors should engage with these communities for outreach. However, we also observe significant usage of social media and conversational AI services for this purpose.


Are you looking to understand where the developers using your product/service go to look for information? Let’s talk. 


About the Author

Nikita Solodkov is a multidisciplinary researcher with a particular interest in using data-driven insights to solve real-world problems. He holds a PhD in physics and has over five years of experience in data analytics and research design.

Comments


Commenting has been turned off.
bottom of page