Guide · · 3 min read
Designing intern projects when AI handles the entry-level tasks
AI now does many tasks interns used to learn on. How to redesign intern projects around judgment, verification, and real decisions.
Jacob Moeckler
Co-founder & CEO
The tasks interns used to learn on, like first-draft research, data cleanup, and meeting summaries, are increasingly done with AI. Employers have noticed: NACE reports that more than one-third of entry-level jobs now require AI skills and nearly 60% of employers assign interns projects using AI tools. Strada found that more than 40% of employers say AI has increased the analytical work given to entry-level employees while reducing routine tasks.
Students aren't all there yet. In NACE's 2026 Student Survey, about half of graduating seniors said they aren't building AI skills. That gap is exactly what a well-designed internship can close.
So design projects around judgment instead of output volume.
Principles that hold up
- Scope for a decision. Every project ends with a recommendation the intern has to defend to a real stakeholder.
- Make AI use explicit. Say whether it's allowed, expected, and how it should be documented. Ambiguity makes strong interns hide good work.
- Evaluate the checking. How did they verify what the tool produced? That's the skill managers will need from them in a full-time role.
- Keep the human skills visible. Strada found employers value critical thinking and communication more highly than AI literacy in entry-level hires. Presentations and stakeholder conversations matter more now, not less.
Before and after
| Old project | Redesigned version |
|---|---|
| Compile a competitive research deck | Use AI to compile it, verify five key claims, and recommend one move |
| Clean a dataset for the team | Clean it with AI, document the checks, and present three findings |
| Draft process documentation | Draft with AI, test it with two real users, and revise |
| Summarize customer interviews | Summarize with AI, then identify what the summary missed |
What this changes in evaluations
If the project is about judgment, the success criteria should be too: quality of the recommendation, how outputs were verified, and how clearly the intern explained their reasoning. Update the evaluation before the summer, not after.
From Jacob: The students I interview are already using AI on the job, and the best examples aren't about producing more. One intern had an AI assistant connected to her company's internal documents and said it cut her ramp-up time because she could get context without scheduling a meeting for every question. Another built an internal AI tool as her actual internship project. A third pointed out the downside: when everyone uses AI to prepare, their work starts to look the same. Judgment is what separates them.
FAQ
Should we restrict AI use for interns?
Set clear rules for sensitive data, then lean toward expected use with documentation. Banning it teaches the wrong habit for the job they're being hired into.
Do managers need guidance too?
Yes. Many managers are still figuring out their own AI use. A one-page project design guide helps more than a policy.
How teams run this in Springbird
Springbird keeps project design and results connected through the offer decision.
- Success criteria in your own evaluation forms, shared at the start so expectations are set on day one.
- Evaluations answered against those same criteria, so judgment and verification show up in the offer decision.
- Resources and tasks, like AI-use guidelines for interns and managers, assigned with completion tracked by name.
- Midpoint feedback scheduled automatically, so a project that isn't working gets fixed while there's still time.
Sources
- NACE: Employer expectations for AI skills meet student skepticism
- Strada Education Foundation: Entry-level hiring in the AI era