Lattice's CEO Gave Three Reasons Companies Are Hiring Juniors Again. Only One Has to Be True

The Wall Street Journal reported this weekend that big employers are hiring again, and that the thaw has reached entry-level roles. We have covered the reversal from several angles already, so the broad strokes are getting familiar.

The new material is testimony from Sarah Franklin, CEO of the HR platform Lattice, who sees across thousands of client companies and says many have gone back into hiring mode for junior positions. Appetite for it, she told the Journal, is considerable.

She also explained why, in three parts.

Companies have realized they need AI-native skills. They want people who are innovative rather than calcified in their thinking. And entry-level workers are "more affordable because they are newer to the workforce."

Franklin is arguing in favor of hiring juniors, and she is right that companies should. But her three reasons describe the same underlying thing, and it is not the thing anybody wants in a press release.

One hire, three documents

Follow those reasons into an actual company and they separate out by document.

The first one goes in the job posting, where AI-native reads as a compliment and works as a filter. What it describes is someone who turned up already fluent in the tools, having become fluent somewhere the company never received an invoice for.

The second one goes in the interview rubric, and it is sincere, and it is also doing something quieter. A candidate with no professional habits to unlearn is a candidate with no settled sense of what this job is worth, how fast the title should move, or what the last three people in the seat were paid before the seat was eliminated.

The third one goes to finance, and it is the only one of the three that has to be true for the requisition to be approved at all.

Line the three documents up and they describe a single person: pre-trained on the expensive new thing at someone else's expense, arriving without the leverage that would let them charge for it, and priced below whoever used to sit there. That is a good deal. Companies are permitted to want good deals, and Franklin is right that this one is worth taking. It is simply not the same event as rediscovering the value of human beings, which is how the week is being narrated, and the whole distinction comes down to who paid for the training.

Somebody else paid for the training

For about two years the entry-level job was what companies cut first and hardest, on the theory that agents could absorb the work juniors used to do. AI-attributed layoffs hit 87,714 in the first five months of 2026, past the whole of 2025, and AI has been the most-cited reason for cuts four months running.

Across those same two years, the people who could not get those jobs learned the tools anyway, because they had no other option. The graduates now being courted for their AI fluency built it on their own time and their own accounts, often while the same employers were explaining publicly that this was the cohort least likely to be needed.

The sequence, then. Cut the roles where the company used to train people. Wait while the labor market trains itself out of necessity. Re-enter and buy that self-funded training at a discount justified by the buyer's inexperience.

Franklin did not design any of that. She is reporting it from the best vantage point available. It still seems worth naming what is being celebrated this week as a return to valuing people.

Fluency is not judgment

A second problem sits underneath the first, and it matters for anyone running a brand function.

We reported recently that AI-native graduates have started confessing online how much they lean on the tools, and that employers are about to inherit that reliance. The confession is the useful part. What these workers have is real fluency: speed, comfort, an instinct for what to ask. What most of them do not have yet is whatever it is that tells you the output is wrong, and that is not a criticism of anyone, because nobody has it at twenty-two. It arrives slowly, through years of being corrected by someone who already had it.

Meanwhile the market is making the opposite bet at the other end of the org chart. Anthropic is paying $300K for a standards editor to argue about commas, and Mercury is paying $335K for a role whose job description reads like an editor-in-chief's. Neither is a vanity hire. They are what judgment costs now that generated output has gone free and somebody still has to separate the good from the merely plausible.

So judgment has become the most expensive thing a company can buy at the same moment that the entry-level hiring wave is being justified on the grounds that the people being hired are cheap. Both things can be true of a market at once. Inside a single company they are supposed to be connected, and the connection is the part worth looking at.

The layer that used to connect them

Nobody hires judgment. Companies manufacture it, and the factory has always sat in the middle of the org chart, in the unglamorous exchange where a junior sends a draft upward and gets it back with the third paragraph circled and a short note explaining why the argument dies there.

The note is the actual product. It carries information that nobody wrote down, usually because nobody could: that this client reads any sentence beginning with "simply" as condescension, that the number is accurate and rhetorically useless, that the company quietly stopped making that particular claim in 2019 after it went badly. Do that a few hundred times over three or four years and the junior stops needing the note, which is the whole mechanism by which a company reproduces its own taste.

That middle layer is considerably thinner than it was two years ago, and the thinning was deliberate. The Fortune 500 spent the year deleting the CMO title, ClickUp is advertising up to $1M for one human to run marketing with three thousand agents underneath and no team at all, and the individual contributor who can build systems is now being handed work that used to belong to a VP. Flattening was the structural fashion of the entire last cycle, and it was sold as a speed argument, which it genuinely was. What left with the layer was the correction mechanism, and that did not appear anywhere in the pitch.

Ford is the version of this story that ends well, which is why it is worth walking through. Jim Farley said last year that AI would replace half of all white-collar workers in the United States. His company then spent three years bringing back 350 veteran engineers to resolve quality problems that automated work had introduced and that nobody remaining in the building could diagnose, and this summer it topped J.D. Power's quality ranking for the first time in sixteen years. Ford now credits engineers working alongside AI rather than AI working instead of engineers. The recovery cost three years and it was only available because the veterans were still reachable, still in the industry, still willing to pick up the phone. A company that runs the same play in 2032 will be looking for people who never received the training in the first place, and it will not find them.

The takeaway for brand leaders

None of this is an argument against hiring juniors. It is an argument for being honest internally about what the hire actually is, because the honesty determines what you have to build around it.

A junior hired on cost and tool fluency is a capacity purchase and it will deliver capacity, which is a real thing to want. What it will not deliver on its own is a brand steward, and the difference between the two is senior time spent on correction. That time is expensive, it competes with everything else on a senior person's calendar, and if it is not explicitly budgeted it will lose that competition every quarter until nobody notices it stopped happening.

There is also a message problem arriving this autumn that will land on marketing rather than HR. The recruitment creative going into market will talk about valuing human judgment and the irreplaceable human perspective, while the stated reason on the investor call is that these workers cost less than the ones who came before them. Both statements are public and permanently searchable. The cohort being recruited is unusually good at spotting that kind of gap, partly because it spent the past two years being told it was surplus by the same companies now courting it, and partly because reading corporate language skeptically is close to a generational reflex.

The discount will also close, and probably faster than the planning assumes. It exists because a generation trained itself during a hiring freeze, and it disappears the moment that cohort has two years of experience and a second offer. Any strategy priced on cheap AI-native labor has roughly eighteen months of runway, after which the labor stops being cheap and the strategy needs a different reason to exist.

Which leaves the least appealing recommendation, the one that costs money before it demonstrates value. The expensive judgment hires that everyone is now reading about, the $300K editors and the rehired veterans, are what a company buys after discovering that it can no longer produce judgment internally. Retaining the people who could have produced it was the cheaper option, and it was available earlier.

Franklin's three reasons are honest, and the honesty is what makes them worth reading closely. She has told anyone paying attention exactly what the market believes it is purchasing, which is a self-trained workforce at a discount, arriving at the precise moment that the mechanism for turning that workforce into something more valuable has been dismantled almost everywhere.

Five years from now this cohort becomes the senior layer. The only correction most of them will ever have received came from a model that was confidently wrong at roughly the same rate they were.