How JBL Uses AI in Recruiting — and Where We Draw the Line

Over the last decade, technology has changed the way we work at JBL Resources in ways we probably would not have predicted when many of our current processes were first built. Search tools have improved, data has become easier to work with, and automation has taken a meaningful amount of manual work out of recruiting.

Like most companies, we have spent a lot of time thinking about where it actually helps, where it simply creates more noise, and where we do not want it involved at all.

Our view of AI in recruiting has become fairly straightforward. We want to use it to reduce the administrative weight that surrounds recruiting, not to reduce the amount of actual recruiting being done. If technology can save a recruiter time on work that is repetitive or procedural, that is useful to us. The value comes from what that recruiter can then do with the time that was saved.

How we currently use AI in recruiting

Much of our use of AI happens behind the scenes. It helps with things like creating backend forms, organizing information within internal processes, summarizing recruiter notes after candidate conversations, and reformatting résumés into a more consistent presentation. These are all necessary parts of the work, but they are not generally the parts where a recruiter's experience or judgment adds the most value.

The same is true on the business development side. We have a large amount of information in our database about companies, contacts, prior conversations, hiring activity, and past relationships. AI can help us sort through that information and identify people who may be worth reaching out to based on what is happening at a company or what we already know about the relationship. In that case, the technology is helping narrow the field. It is not deciding what the relationship should look like or carrying it forward on its own.

That distinction is important to us because the more capable these tools become, the easier it is to automate things simply because they can be automated. We've tried to resist that instinct. Some parts of recruiting are inefficient for a reason: they involve people, context, incomplete information, and judgment. Those are not always problems that need to be removed from the process.

What we want technology in staffing to give back

When we think about technology in staffing, the most useful question for us is not how much more activity it can generate. It is how much meaningful time it can give back to the people doing the work.

A recruiter may finish a candidate conversation with pages of notes that need to be organized. If AI can reduce the time spent cleaning those notes up, that is valuable. If a résumé needs to be reformatted before it can be sent to a client, there is little benefit in having someone spend 20 minutes manually correcting spacing and headings. If we are trying to determine which contacts in our database are most relevant to an active hiring need, there is no reason to ask a recruiter to comb through hundreds of records one by one.

In each of those cases, we are comfortable allowing technology to do more of the administrative work because it leaves more room for the parts of recruiting that require a person. That might mean having another conversation with a candidate, asking a hiring manager a better follow-up question, or spending more time understanding why someone's experience does or does not fit an opportunity.

That has gradually become the standard we use when looking at AI automation in recruiting. If the technology creates more time for judgment and conversation, it is probably helping. If it starts replacing those things, we become much more cautious.

Where we have chosen not to automate

There are several areas where we have decided that the efficiency gain is not worth what would be lost.

One is recruiter outreach. AI can help identify who may be worth contacting, but we do not want it writing and sending emails in place of the recruiter. Outreach is the beginning of a relationship, and we believe the person whose name is on the message should actually be the person behind it. That means taking the time to understand why someone is being contacted and writing the message accordingly.

We take a similar view of candidate conversations. There are tools available that can automate more of the screening process or replace parts of an interview with an AI-driven interaction. That is not a direction we are interested in pursuing. A résumé gives us useful information, but a conversation often changes how we understand it. Candidates explain decisions, add context, ask questions, and describe experiences that do not always translate cleanly to a document or database field. The human connection in hiring is not separate from the evaluation. In many cases, it is how the evaluation becomes accurate.

The clearest boundary for us is candidate rejection. We will not allow AI to automatically reject a candidate. Technology can help organize information, identify areas that deserve attention, or compare a résumé against known requirements. The final judgment belongs to a person.

Part of the reason is practical. Recruiting is full of imperfect matches that turn out to be very good ones. Job titles differ between companies. Skills transfer between industries. People take unconventional career paths. A candidate who looks slightly off-target on paper may make complete sense after a conversation. We do not want to build a process that becomes more efficient by removing the opportunity to notice those things.

There is also a broader principle behind it. As AI becomes more capable, keeping a human in the loop will require more deliberate choices, not fewer. It will often be possible to automate another step. The question for us is whether doing so actually improves the quality of the work.

We expect our use of AI to continue growing. There are almost certainly administrative tasks we do manually today that we will not be doing manually a few years from now. We see that as a positive development. The part we are more interested in protecting is the time that sits on the other side of that efficiency.

If a recruiter spends less time formatting documents, organizing notes, or searching through records, we would like that time to show up somewhere else in the process as a better conversation, a more thoughtful evaluation, or a stronger relationship. That is the balance we are trying to maintain as the technology changes, and it is likely to remain the way we think about AI for some time.