Shilpi Sardana, Senior Director – India Operations & People Strategy at Netomi
By Shilpi Sardana, Senior Director – India Operations & People Strategy at Netomi
There’s a growing consensus in HR circles: AI is reshaping jobs faster than anyone can rewrite a job description. The Software Engineer is becoming an AI Software Engineer. The Product Manager is becoming an Autonomous Agent Product Manager. The recruiter is becoming an AI Talent Advisor.
That part of the story is right. But there’s a critical piece missing from most of this conversation, and it changes everything about how HR should respond.
You can’t hire your way out of this. The talent you need doesn’t exist yet – not at scale, not in the market.
The role that’s transforming
Forget the abstractions for a second. Look at one real job: the Customer Success Manager.
The traditional CSM, relationship owner, renewal shepherd, QBR host is becoming obsolete. Not because customer success stopped mattering, but because AI products don’t need a relationship manager nearly as much as they need someone who can build. Someone who can configure the product, design the workflow, debug the integration, and sit with the customer’s engineering team to make the thing actually work.
That job already has names:
Solutions Engineer, technical depth plus customer fluency, the person who architects the implementation, not just manages the account.
Forward Deployed Engineer, embedded with the customer, hands-on, building custom solutions in the field rather than shipping generic playbooks.
Prompt Engineer, not a job title that existed three years ago, now a baseline skill for anyone shaping how AI systems behave.
None of these are CSM with a new business card. They’re a different job, requiring a different kind of person.
Why “just hire for it” doesn’t work
Here’s the problem every HR leader is about to run into: you cannot recruit your way into roles that don’t have a resource market yet.
A Forward Deployed Engineer with five years of experience doing exactly this job doesn’t really exist, because the role didn’t exist five years ago. The handful of people who’ve done it are either already deeply embedded in the handful of companies that invented the role, or they’re commanding compensation that makes “just go hire some” a fantasy for most organizations.
Waiting for the market to produce this talent is a losing strategy. By the time it does, the companies that built it internally will already be three years ahead.
The answer isn’t build or buy. It’s build and buy — deliberately
This isn’t an argument against external hiring. It’s an argument against treating external hiring as the default strategy, when for these specific roles it can’t be.
Buy, sparingly and strategically. Bring in a small number of external anchors who’ve done this at another company, even a different industry. Their job isn’t to fill headcount — it’s to compress the learning curve for everyone else and set a credible bar for what “good” looks like.
Build, deliberately and at scale. This is where the real capacity has to come from. Look inside first:
● Your CSMs who already understand the customer’s problems better than any outside hire ever will — give them the technical skills to become Solutions Engineers.
● Your QA engineers who already think in edge cases and failure modes – that’s exactly the mindset AI evaluation work needs.
● Your technical writers who can already translate complexity into clarity – that’s most of what prompt design actually is.
The raw material for these new roles is already on your payroll. What’s missing isn’t potential. It’s a reskilling path that takes it seriously.
The one thing worth measuring before anything else
If you only do one thing before building a reskilling program, do this: figure out who on your team actually has the learning agility to make this jump, independent of how long they’ve been in their current role or how they’ve performed against it historically.
Tenure and past performance tell you who was good at yesterday’s job. They tell you almost nothing about who can become tomorrow’s. That’s a different question: and most performance systems were never built to answer it.
This is exactly why performance and potential need to be evaluated separately, not blended into one score. A steady, dependable CSM and a CSM who’s quietly been teaching themselves SQL on weekends can look identical on a performance scorecard. They are not identical candidates for a Solutions Engineer reskilling track and the only way to tell them apart is to have deliberately asked the potential question, not just the performance one.
To identify this “learning agility,” look for these three behaviors:
● Edge-case curiosity: They habitually ask “what if” about product features.
● Self-directed experimentation: They are actively using AI tools in their personal workflows or side projects.
● Rapid prototype iteration: They show a pattern of testing, failing, and adjusting rather than waiting for a perfect plan.
Where this leaves HR
The organizations that win the next five years won’t be the ones with the best recruiting funnel. They’ll be the ones that got two things right at the same time:
● Identified early which internal roles are heading toward obsolescence, and which adjacent roles are already emerging to replace them.
● Built the reskilling infrastructure to move people deliberately from one to the other – before the market forces a slower, more expensive, more disruptive version of the same transition.
Hiring will still matter. But it stops being the main lever. The main lever is figuring out, faster than your competitors, who on your existing team can become who your company needs next – and then actually investing in getting them there.
The skills gap is real. But it’s not a gap the market is going to close for you. It’s one every organization is going to have to build its way across, on its own timeline, with its own people.
Immediate Next Steps
Audit Roles: Identify five internal roles currently most vulnerable to AI shifts.
Pilot a Path: Launch a reskilling track for ten high-potential employees to transition into an adjacent technical role.
Update Metrics: Separate ‘potential’ from ‘performance’ in your next review cycle to surface hidden talent.
(Views expressed in the article are solely of the author.)
About The Author
