AI for IT teams: which tasks to automate first
Updated
Across the 5.5 million US jobs in IT and software, the work scores 58 of 100 for AI exposure, weighted by how many people do each job. AI can take on about 892 hours a year per full-time person, 45% of a working year. The task that frees the most hours across the team is to prepare reports or correspondence concerning project specifications, about 114 hours a year for each person who does it.
- AI exposure, weighted by employment
- 58 out of 100 High exposure
- Hours AI can take on
- 892 hours a year per full-time person
- Task time AI can do today
- 12% AI speeds up another 78%
Tasks to automate first
The 15 tasks where AI can take on the most hours across IT and software: how exposed each task is, how much of the job it takes and how many people do that job. Hours are per person in the role that does the task. How we score each task.
| Task and who does it | AI exposure |
|---|---|
| Prepare reports or correspondence concerning project specifications, activities, or status. Status reports and specification write-ups are routine writing AI handles well. Done by software developers. | 78% AI can do it 114 hours a year for each person in the role |
| Develop or direct software system testing or validation procedures, programming, or documentation. AI drafts test cases, validation scripts and docs; engineer directs and approves. Done by software developers. | 65% AI helps 101 hours a year for each person in the role |
| Store, retrieve, and manipulate data for analysis of system capabilities and requirements. Querying, transforming and retrieving data is routine digital work AI scripts handle. Done by software developers. | 68% AI can do it 97 hours a year for each person in the role |
| Maintain records of daily data communication transactions, problems and remedial actions taken, or installation activities. Logging transactions and remedial actions is largely templated and automatable. Done by IT support specialists. | 72% AI can do it 98 hours a year for each person in the role |
| Answer user inquiries regarding computer software or hardware operation to resolve problems. LLM assistants answer most routine software and hardware questions. Done by IT support specialists. | 65% AI can do it 97 hours a year for each person in the role |
| Read technical manuals, confer with users, or conduct computer diagnostics to investigate and resolve problems or to provide technical assistance and support. AI speeds troubleshooting and manual lookup; live user contact remains. Done by IT support specialists. | 55% AI helps 85 hours a year for each person in the role |
| Prepare and review operational reports or project progress reports. AI can draft progress reports from project data for review. Done by IT managers. | 70% AI can do it 80 hours a year for each person in the role |
| Stay abreast of advances in technology. AI summarizes tech news and research rapidly for review. Done by IT managers. | 65% AI helps 80 hours a year for each person in the role |
| Provide users with technical support for computer problems. AI help desk tools resolve much routine user support. Done by IT managers. | 60% AI helps 76 hours a year for each person in the role |
| Use object-oriented programming languages, as well as client and server applications development processes and multimedia and Internet technology. AI writes and reviews most application code; developer is responsible for design. Done by systems analysts. | 60% AI helps 63 hours a year for each person in the role |
| Use the computer in the analysis and solution of business problems, such as development of integrated production and inventory control and cost analysis systems. AI builds cost models and analyses; analyst judges business fit. Done by systems analysts. | 55% AI helps 62 hours a year for each person in the role |
| Troubleshoot program and system malfunctions to restore normal functioning. AI parses logs and suggests fixes; engineer verifies live recovery. Done by systems analysts. | 50% AI helps 60 hours a year for each person in the role |
| Document computer security and emergency measures policies, procedures, and tests. Drafting policies, procedures and test reports is highly suited to LLMs. Done by cybersecurity analysts. | 75% AI can do it 133 hours a year for each person in the role |
| Modify computer security files to incorporate new software, correct errors, or change individual access status. AI can script access and config changes; reviewers approve before applying. Done by cybersecurity analysts. | 62% AI can do it 114 hours a year for each person in the role |
| Maintain logs related to network functions, as well as maintenance and repair records. Log and repair records are documentation AI handles well. Done by systems administrators. | 70% AI can do it 69 hours a year for each person in the role |
What stays with people
10% of the task time across these roles needs a person, and the output of the rest still needs someone to check it. The tasks below take the most of that time.
- 12% AI can do it
- 78% AI helps
- 10% Needs a person
- Supervise the work of programmers, technologists and technicians and other engineering and scientific personnel. Software developers, 15% exposed
- Set up equipment for employee use, performing or ensuring proper installation of cables, operating systems, or appropriate software. IT support specialists, 15% exposed
- Install and perform minor repairs to hardware, software, or peripheral equipment, following design or installation specifications. IT support specialists, 18% exposed
- Supervise and assign work to programmers, designers, technologists, technicians, or other engineering or scientific personnel. Software developers, 14% exposed
- Meet with department heads, managers, supervisors, vendors, and others, to solicit cooperation and resolve problems. IT managers, 10% exposed
Adoption path
Start with the tasks where AI helps a person, then automate the routine ones, then give whole workflows to agents with a person approving the output.
Roles in IT and software
27 roles, each linked to its task-by-task score. Of the roles with at least 2% of the jobs, software testers score highest at 64, IT managers lowest at 50.
- Software developers, 30.9% of the team, score 59 out of 100
- IT support specialists, 13.1% of the team, score 59 out of 100
- IT managers, 12.3% of the team, score 50 out of 100
- Systems analysts, 9.5% of the team, score 59 out of 100
- Systems administrators, 5.8% of the team, score 58 out of 100
- Cybersecurity analysts, 3.5% of the team, score 60 out of 100
- Software testers, 3.4% of the team, score 64 out of 100
- Computer network architects, 3.3% of the team, score 55 out of 100
- Network engineers, 2.7% of the team, score 59 out of 100
- Programmers, 1.7% of the team, score 66 out of 100
- Computer hardware engineers, 1.4% of the team, score 51 out of 100
- Web developers, 1.3% of the team, score 66 out of 100
- Database administrators, 1.3% of the team, score 60 out of 100
- Data engineers, 1.2% of the team, score 60 out of 100
- Blockchain engineers, 0.9% of the team, score 58 out of 100
- Digital forensics analysts, 0.9% of the team, score 59 out of 100
- Document management specialists, 0.9% of the team, score 61 out of 100
- GIS technicians, 0.9% of the team, score 62 out of 100
- Information security engineers, 0.9% of the team, score 60 out of 100
- Information technology project managers, 0.9% of the team, score 56 out of 100
- Penetration testers, 0.9% of the team, score 60 out of 100
- Systems engineers, 0.9% of the team, score 60 out of 100
- Web administrators, 0.9% of the team, score 62 out of 100
- Computer scientists, 0.7% of the team, score 51 out of 100
- Data warehousing specialists, Counted under a related job in US data, score 68 out of 100
- Health informatics specialists, Counted under a related job in US data, score 58 out of 100
- Telecommunications engineering specialists, Counted under a related job in US data, score 57 out of 100
Questions about AI for IT teams
Which tasks should an IT team automate first?
The first task to automate is to prepare reports or correspondence concerning project specifications (78% exposed). Next come the tasks to develop or direct software system testing or validation procedures (65%) and to store, retrieve, and manipulate data for analysis of system capabilities and requirements (68%). Together they free the most hours across the team.
How many hours can AI save an IT team?
About 892 hours a year per full-time person, 45% of a 2,000-hour working year. For a team of 10 that is about 8,920 hours, the time of 4.5 full-time people. The figure is the share of task time AI can remove today, weighted by how many people do each job.
Which roles in an IT team are most exposed to AI?
Of the roles with at least 2% of the jobs, software testers score highest, at 64 of 100, and IT managers lowest, at 50. Each role's page lists its tasks one by one.
Will AI replace jobs in IT and software?
AI can do 12% of the task time across these roles today. Another 78% gets faster with AI while a person stays in charge. 10% needs a person. One task that stays with people is to supervise the work of programmers.
Where should an IT team start with AI?
With the quick wins: approved assistants for the tasks AI helps with, under a written AI policy. Then automate the 8 tasks on this page AI can do end to end. An AI audit maps the tasks in your team and ranks them the same way this page does.
Next steps
- Enterprise AI in Slack: one assistant for company knowledge
- Case study: shared code context for a development team
Find the tasks to automate first in your team
We map the tasks in your IT team, score each one the way this page does and build the automations that pay back first.