The new grad gold rush isn't in AI engineering — it's in making AI actually work inside companies that have no idea what to do with it.
The Summary
- Jiaona Zhang, CPO at Laurel and Stanford lecturer, says "AI workflows" is the hot entry-level role — identifying where AI can optimize company operations and implementing those changes
- One Laurel new grad built an agent that acts as a personal chief of staff for salespeople, becoming "the most celebrated person at this company"
- Zhang calls it "the new Biz Ops" and says grads should create this role at companies even if it doesn't officially exist yet
The Signal
Here's what Zhang is really describing: the gap between having AI tools and knowing what the hell to do with them. Companies are drowning in ChatGPT seats and agent platforms. Most employees use them like fancy search engines. Someone needs to translate "we have AI" into "here's the automated workflow that just saved sales 40 hours a week."
The AI workflows role sits between technical and operational. You don't need to build the LLM. You need to see that your sales team spends three hours a day writing follow-up emails and another two scheduling demos, then spec out an agent that does both while the humans do discovery calls.
"Instead of being the single salesperson hitting your quota, you're able to scale your impact across the entire sales team."
This is the arbitrage opportunity for new grads right now:
- Companies have budget for AI tooling but no internal expertise on implementation
- Mid-career employees are intimidated by agents or too buried in existing work to experiment
- Fresh grads who've been using Claude and GPT since sophomore year see automation opportunities everywhere
The Laurel example is instructive. A new hire built a chief of staff agent for salespeople. Not a chatbot. Not a dashboard. An agent that actually handles the administrative load that keeps salespeople from selling. That person became the most valuable non-engineer in the building because they made everyone else more effective.
Zhang frames this as "AI Ops" parallel to Business Operations, the catch-all function that emerged in the 2010s for "figure out how this company actually runs and make it run better." Biz Ops people became indispensable because they understood both the business and the tools. AI workflows is the same play for the agent economy.
Key differences from traditional ops roles:
- You're not just optimizing processes, you're rebuilding them around agents
- The tooling changes every quarter, so you need to stay current on what's possible
- You're often creating ROI that's immediately visible (hours saved, tasks automated, bottlenecks removed)
The tactical move Zhang suggests: don't wait for the role to exist. If you're in sales, build the automation. If you're in customer success, spec the agent that triages tickets. Prove the value, then either get promoted into the formal role or take that track record somewhere that will pay for it.
This is classic early-market behavior. In 2008, companies didn't have "social media manager" roles until someone in marketing started running the Twitter account and drove actual results. Same pattern here. The role crystallizes after someone proves it works.
The Implication
If you're a new grad, this is your wedge. You don't need a CS degree. You need pattern recognition for where work is repetitive and annoying, plus enough technical literacy to use agent platforms and automation tools. Learn to prompt well, understand API basics, and get good at translating "this sucks" into "here's the agent that fixes it."
For companies: if you're hiring new grads and not carving out space for this function, you're letting them go somewhere that will. The ROI on someone who can make your whole sales team 30% more efficient is higher than the ROI on one more seller.