INFOGRAPHIC

Insights From the AI Frontier

Workers involved in AI development think they still have a lot to learn as they navigate this technological frontier. Meanwhile, agentic AI is getting widespread attention as the trustworthiness of models looms large as a top concern. These were among the findings of a recent survey of enterprise AI developers. Commissioned by IBM, Morning Consult polled 1,063 enterprise AI developers in the fall of 2024. 

 

To qualify for the survey, respondents were required to fill one of the following roles: data scientist, application developer, system developer, AI developer, machine learning (ML) engineer, software engineer, software developer, AI engineer or IT engineer. They also had to at least occasionally contribute to the development of enterprise AI applications in their current role. 

 

Following is a selection of the survey’s findings.

Source: IBM, "Survey: Generative AI Makes Tasks Simple, But Developing That AI is Anything But," January 2025.

Expert Experience Level With GenAI

Expertise in Generative AI (GenAI) varies widely among developers. 
0
%
AI Developer
0
%

Data Scientist

0
%
Software Engineer
0
%
System Developer
0
%
ML Engineer
0
%
Software Developer

Experience Level With Types of AI

Developers have more experience with GenAI than classical AI.

0
%

Expert in GenAI

 

0
%

Expert in classical AI

Time Saved Per Day Using AI-Assisted Coding Tools

AI developers are getting a hand from pre-existing AI tools, as AI coding assistants prove to be a significant time saver.
0
%

Less than 15 minutes

0
%

15-29 minutes

0
%
30-59 minutes
0
%
1-2 hours
0
%
3-4 hours
0
%
Over 4 hours

Types of Development Tools, Extent of Usage by Level 

Over half of AI developers are using no-code (requires no traditional coding) or low-code (requires some traditional coding) development tools.

0
%

Pro code (traditional coding)

 

0
%

Low code

0
%

No code

Top Challenges of Developing GenAI at the Enterprise Level

AI trustworthiness and the lack of a standardized AI development process topped the list of challenges faced by developers.
0
%

Lack of a standardized AI development process

0
%

Ethics, trust and data transparency and traceability

0
%
Customization
0
%
Rate of Change
0
%
Infrastructure/stack complexity
0
%
Establishing governance and ensuring compliance

Top Concerns for Scaling AI Agents in Enterprise 

99% of enterprise AI developers are exploring or formulating use cases for AI agents, and as they scale those tools, their top concern is trust. 

  • Trustworthiness (ensuring outputs are accurate and void of bias): 31%
  • Introducing new attack vectors (AI agents being compromised by malicious actors): 23%
  • Adhering to compliance and regulations: 22%
  • Becoming overly autonomous (humans losing oversight and visibility into systems): 22%
  • No concerns: 3%

31%

Trustworthiness (ensuring outputs are accurate and void of bias)

23%

Introducing new attack vectors (AI agents being compromised by malicious actors)

22%

Adhering to compliance and regulations

22%

Becoming overly autonomous (humans losing oversight and visibility into systems)

3%

No concerns

AI Agent Use Cases Being Explored

Top use cases for agents include customer support, project management and content creation.
0
%
Customer Service and Support
0
%
Project Management / Personal Assistant
0
%
Content Creation
0
%
HR
0
%
Transportation
0
%
Health Care
Share this article
HOME