You were hired for what was on your resume. You will be kept, promoted, or quietly managed out for something that was not listed on it.
That gap is the whole problem with soft-skills hiring, and it is why Sameer Ranjan built an AI platform to close it. He started as a mining engineer, graduated at the top of his class, then walked away from the field entirely to build talent systems in Dallas. He holds a US patent for the platform that resulted from it. When I asked him how long resumes have left, he did not hedge. Eighteen months, on the outside.
I have some skin in this. A few years ago, I stopped reading CVs before interviews. If a recruiter had already screened for credentials, reading the document again only told me what to be biased about. I wanted to meet the person, not the paper. So this conversation landed somewhere personal.
Soft skills hiring is the gap: companies hire for job skills and lose people over soft skills
Sameer put it plainly. People get hired on day one for their job skills. They stay or they leave because of their soft skills and their personality.
Nobody survives twenty years at a company because they were an excellent coder. They survived because they were a good team player, a decent manager, and someone who showed up when it was inconvenient. And almost none of that gets assessed in an interview. The board has no idea why one person stays for two decades and another leaves in two years, because the thing that decides it was never measured.
This is a measurement issue, not a people issue. And it maps directly onto the Ownership vector in the 4 Productivity Vectors methodology. Accountability, support systems, and cross-functional work are exactly the capacities that decide whether someone lasts, and exactly the capacities a resume is silent about.
The three failure points in how companies hire today
Sameer sees three, and they compound.
The first is that companies do not know what they need. If you cannot define the need, you either overhire, hire the wrong profile, or hire too few. All three have happened across the market since 2023. Companies overhired, then cut. Then they froze because nobody could say who was actually required.
The second is a confusion about what AI replaces. Leaders look at a team of six and assume AI takes five of those jobs. It might. It might also make those six people substantially better at their work. Deciding which one you are betting on changes the entire hiring plan, and most companies have not made that decision explicitly.
The third is speed. Recruiters want to process more resumes and more candidates, so they delegate the job description to a chatbot and the interview to an AI tool, then pick from the top of the list. What comes out the other end is a shortlist of people who are not culturally fit for the organization.
Why an AI interview cannot do the one thing interviews are for
I have been vocal about AI interviews, and I am against them. Sameer gave me the clearest explanation I have heard for why.
Tell a human interviewer that someone in your family died, and they feel it. They relate it to something in their own Life. That reaction is not a nicety; it is the mechanism by which humans read each other. A machine can register the fact. It can produce sympathy. It cannot produce empathy, because empathy requires having been there.
Sameer draws the line. Sympathy is feeling bad when someone falls in the street. Empathy is walking over, helping them up, and cleaning the wound. One is a feeling. The other is a decision followed by an action. If the thing assessing you cannot make that distinction, it cannot assess the part of you that actually predicts how you will work with other people.
Interviewers are worse at reading people after 1pm
Here is the detail that should make every hiring manager uncomfortable. Sameer’s data shows that interviewers are often wrong about the people in front of them, and that accuracy depends on the time of day.
Someone who started interviewing at 8am is running on fumes by early afternoon. Meetings are more tiring than the work itself. Fatigue clouds judgment, and that judgment is about someone’s career.
Then there is the obvious problem with the format itself. Anyone can be charming for thirty minutes and be a completely different person for the other twenty-three and a half hours. A half-hour sample is not a read on a personality. It reads like a guess rather than a process. This is a Decision-making problem inside the Effectiveness vector, and most organizations have never treated it as one.
Soft skills hiring in practice: what MayaMaya measures
MayaMaya is built on behavioral economics, drawing on Daniel Kahneman’s work in Thinking, Fast and Slow and Noise. Instead of asking you to rate yourself, the assessment puts you in everyday situations and asks you to arrange what you would do and in what order. There is no single correct answer, and your answer can change depending on the week you are having.
That variability is the point. Personality and soft skills are shaped by what you are living through, so a single snapshot is close to useless. The platform runs the assessment on a repeating cycle, quarterly or twice a year, to establish a baseline and then track the movement around it.
Combine that with productivity data and resume data, and you get what Sameer calls a digital twin of a person. From there, a company can ask better questions. Will this person work on this team? Will they still be here in three years? Can they move into a different role? Are they ready to lead?
Personality drift is an early warning system for burnout
The continuous tracking has a use that has nothing to do with hiring, and it is the part I find most interesting.
People change. Someone gets married and shows up differently. Someone loses a family member and changes. That is normal, and it registers in behavior. Once you have a baseline, you can see how much it deviates.
Sameer’s argument is that a large share of people who quit are leaving because of burnout, and burnout shows up in personality data before it shows up in a resignation letter. If you can see the drift, you can do a root cause analysis and intervene while intervention still means something. That is the Well-being vector operating at organizational scale rather than as an individual’s private problem. If you want to see where your own load is concentrated, the productivity quiz is a reasonable place to start.
The privacy line, and who owns the data
I pushed him on this because the same system that helps a company support you can be used to watch you. Personality data is about as personal as workplace data gets.
Two design choices carry the weight. The first is data sovereignty. The employee holds the access. They decide whether the organization sees it, and they can switch that off. Sameer describes it as a data passport that belongs to the person, not the employer.
The second is that the platform is positioned as decision-aiding rather than decision-making. It does not tell a company to fire someone. It presents the reasoning and the options and leaves the call with a human. Whether that boundary holds under commercial pressure is a fair question, and the answer will vary by the company operating it.
How to audit your own soft skills this week
Sameer’s method is very straightforward. Go to a restaurant and start a conversation with a stranger.
You do not need to be an extrovert. Please go out and talk to real people, listen properly, and make someone feel heard. That is the only place a soft skill can actually be validated, because it only exists in contact with another person. If the conversation goes nowhere, you have learned something useful about where to work.
If you want a broader diagnostic on why capable people stall or leave, I wrote about it in “Why High Performers Quit.” He is blunt about why this matters more now. A generation entered the workforce during remote work and never built the reps. The skill that decides how long you last at a company is the one they had the fewest chances to practice.
Different industries need different soft skills
The patterns in Sameer’s data are specific enough to be useful.
Healthcare and nursing run on empathy paired with clear communication. Using the Ikigai lens of spirit, profession, passion, and reward, the healthcare people scored high on spirit and very high on professionalism, and notably low on reward.
IT and manufacturing run on negotiation, and not the money kind. Negotiating ideas. Being able to accept someone else’s idea and get yours accepted. IT profiles came out high on spirit, lower on professionalism, very high on passion because they are problem solvers, and high on reward outlook because they want to be paid well for it.
Sales runs on leadership presence. Commanding a room, then reading accurately whether you got the deal after you leave it. Sales profiles scored high on reward, passion, and purpose because the purpose is singular.
None of this is a cage. Sameer is emphatic that you can upskill into a profile you do not currently have. The value is knowing which direction you are starting from.
Why does the resume have about eighteen months left
His reasoning is straightforward. Within roughly eighteen months, the baseline technical skills level out. Everyone will know how to prompt. Everyone will be able to assemble code with a tool that spins up the architecture for them. When the base layer is common across every candidate, listing it stops being a differentiator.
What is left is the third layer. How do you solve a problem? How fast you learn. How adaptable you are. How you communicate an idea and whether you actually listen. Those are the hiring criteria that survive, and they are the ones a document cannot demonstrate.
He carves out an exception, and it is a fair one. Trades still need proof of work. You cannot hand someone a spanner and expect them to fix an HVAC system. But for accountants, analysts, engineers, marketers, and salespeople, he expects the resume to become obsolete.
Organizations are measuring the wrong kind of productivity
This is where the conversation connected back to what I work on every day.
Most companies take whatever their ERP system reports and treat it as the truth. But if the scope of the work changed mid-quarter and the system never recorded the change, productivity would read as collapsed, even though every shift was running seven days a week. The number is not measuring output. It is measuring how the data was collected.
It gets stranger with AI in the mix. Please assign ten employees to ten agents, and ensure every task closes ahead of schedule. On the dashboard, productivity is perfect. But did any of those ten people gain a skill they did not have before? Nothing in the reporting answers that.
Sameer’s alternative is to measure growth alongside output. If every person in the organization gained two percent in capability over a period, the organization is compounding in a way no throughput metric captures. That is a granular, employee-level measure rather than a ten-thousand-foot one, and it is the difference between an organization that is busy and an organization that is getting better. The same distinction I make with individuals is applied to a company.
The one thing to take from this
The skill that keeps you employed is one nobody has been measuring, and it exists only in contact with other people. So practice it in contact with other people.
This week, have one conversation you would normally avoid. A colleague in another function, a stranger at a counter, the person on your team you have never spoken to outside a status call. Listen more than you talk. That is the audit, and it costs nothing.
More on building the habits underneath all of this sits in the productivity hub. Sameer’s last word on where people go wrong with AI at work is worth keeping, too. They treat it as a search engine when it is an assistant. Those are two very different relationships, and only one of them makes you better at your job.
Want to know where your own productivity is leaking before someone else measures it for you? Take the free Productivity Assessment.