Posts
Hesitation is the last honest signal
See other posts from by Bryan Adams
CEO
When an AI can write and submit a thousand applications before breakfast, the only data left worth trusting is the moment a real human pauses.
“If you’re not clear on your own story, it’s almost impossible to tell someone else’s truth with clarity.” Seth Godin, in conversation with me
Years ago, I ran a workshop using a simple story structure to warm people up. I asked employees to share a career moment when they faced adversity and pushed through. The stories came one after another, and they all technically worked. You could tick every box. And every one of them sounded like it could have come from any company on earth. So, I asked one woman if I could workshop her example live. I did not change her facts. I just kept asking what was really at stake, when the turning point hit, and how she felt afterwards. The story deepened. The room leaned in. Afterwards she told me the new version felt closer to what had really happened, not because anything had changed, but because she finally had space to express what it meant. The facts were never the signal. The meaning underneath them was. And the meaning only showed up when I stopped counting and started reading.
Hold that distinction, because hiring is about to need it more than it ever has.
Let us say the quiet part loudly. The job application just died as a meaningful signal, and AI was only the last nail. It was always soft. Rising application volume never meant rising quality, it meant you had reduced friction or bought more reach. We liked the number because it was easy to count and looked good in a board deck. Now agentic AI will tailor and fire off a thousand applications on a candidate’s behalf while they sleep. The funnel is not leaking. It has been flooded with water that looks exactly like signal and contains none.
Half the industry is responding by building better filters, more AI to fight the AI, an arms race that ends with machines talking to machines and humans nowhere in the conversation. That is the wrong war. Here is the shift instead.
In a world where action is free and infinite, the only un-fakeable signal left is human hesitation. A bot can apply. A bot cannot care. It does not return to a role page three times over four days because something is nagging at it. It does not read the culture page, then benefits, then circle back to the role, quietly assembling a picture. It does not dwell. It does not waver. Hesitation is the one thing in the entire process that AI cannot counterfeit, because it is the visible residue of a real person weighing a real decision.
This is not even a new problem. It is a borrowed solution we never picked up. Fifteen years ago, B2B marketing had the identical crisis. A lead stopped meaning anything because forms were free to fill. Their answer was not to count leads harder. It was intent data: reading the pre-purchase behavior. Who came back. Who lingered on pricing. Who consumed three things before raising a hand. An entire discipline grew out of that one reframe. Recruiting is sitting exactly where marketing sat in 2010, staring at a vanity metric and wondering why it stopped predicting anything.
The pre-apply behaviors are already there, and they sort cleanly into five families worth reading: attention, what candidates notice; consideration, what they are trying to understand; confidence, what nudges them closer to acting; friction, where interest stalls; and quality, which journeys produce the better conversations, not just the clicks. Notice what they have in common. The most valuable candidate signal in your entire process is generated by people who have not formally entered your process yet. We have been staring at the finish line and ignoring the whole race.
Putting our money where my mouth is
We made an early, deliberate bet at Happydance that pre-apply behavior, not application volume, is where the next decade of candidate-experience advantage gets built. So the platform is instrumented to read the hesitation, not only the conversion job views, saves, and shares against apply clicks; applications started against applications completed; what candidates search for, how they refine it, and which journeys end in action.
“Our conversational navigation goes further: it tells customers what candidates are actually asking, intent that traffic data alone cannot see. .
On the roadmap we are building an AI analytics layer that does the reading for you: automated detection of drop-off points, underperforming pages, and content gaps, translated into plain-language priorities rather than charts. It is a bet that the next advantage is not collecting hesitation data but understanding it faster than your competitors do.
National Grid saw it firsthand. Search behavior on their careers website revealed candidates actively hunting for apprenticeship and intern roles, demand their traffic analytics alone could never have surfaced, and content gaps were closed in response. The same instrumented site converts job views to apply clicks at 13%, feeding candidates directly into their ATS.” Jim Taylor, MD & COO, Happydance
We spent ten years engineering every reason to slow down out of the candidate journey, and now we are surprised we cannot tell who really wants the job. Applications tell you who acted. Hesitation tells you who cared. In the age of AI, only one of those is still true.
