The Parable of the Blind Men and the Elephant: Task Automation and the Emergence of Workforce Intelligence

In record time, the whole HCM space has converged on the idea that work itself is changing. And that the atomic unit of work is the task. Tasks are now recognized as a critical unit of change – it’s where automation happens. It’s therefore the driver of a brand new question we need to ask: “not just ‘who’ should do this work, but ‘what’ should do this work?” Robots, agents, and GenAI are now part of the workforce. But no one seems to be asking “once we know the who or what – what downstream activities does this set in motion...?”

For the last two years, a new concept has been emerging, and it’s now a widely accepted "new normal": in an automation-driven world, the unit of change is the task, not the job. Automation doesn't retire a role or a process cleanly. It goes after the tasks – augments some, kills others, and hands the role or process back to us looking nothing like the job description or workflow on file.

The good news is that we all reached this conclusion in record time.  From “skills-first strategies” to “task first + skills” in something like 18 months.  The bad news is that most organizations are still in the analysis phase – which tasks, role, and processes have the greatest automation exposure? But here’s the thing – that’s actually the easy part.  Every vendor worth their salt can do automation exposure assessments. The hard part is what comes next.

Knowing or unknowingly, once we all acknowledged that the work is changing at an atomic level, we actually signed up for four more downstream projects – each of them enormous in their own right.

Work Redesign

When the tasks change, the work has to be redesigned – which tasks stay human, which get augmented, which get automated away.  Role redesign, process redesign, skills-based hiring strategies – all of these need to become part of the DNA of every HR team.

Continuous, Always-on Job and Work Architecture Redesign

When the work is redesigned, the jobs and the work architecture need to change; roles assembled from the old task mix no longer describe what people actually do all day, nor do they reflect the new and emerging skills required, or the skills obviated by automation impacts. And this isn’t going to be a “one and done” exercise – this will be a new review every six months for the next 10 years.  Which is a problem since the average WTW or Korn Ferry job architecture effort takes 12 months (best case) to complete and is out-of-date on the day it launches.

Workforce Redesign and Reskilling

When the jobs and the work architecture are redesigned, the workforce has to change – the reskilling, the upskilling, the redeployment the new task mix demands.  What percentage of the people who are redundant or augmented by automation can be repurposed?  To what roles?  Should we hire people into roles with high automation exposure, or should we contract since we don’t know what’s happening with the work?  All new questions requiring new tooling and deep workforce data from both inside and outside the organization.

Intelligent Organization Design and Redesign

And when the workforce changes, the organization has to change because a structure built to hold the old roles doesn't fit the new ones.  If x% of this organization goes to automation and y% of this other organization does the same, how adjacent are the remaining tasks?  Can we merge the orgs?  What if we deploy a set of agents and rethink workflows that normally span multiple teams – can we collapse the walls between these teams?  What is the right span of control for an AI-augmented team?

Work. Jobs. Workforce. Org. That's the chain. Pull on the first link and all four move.

Now here's the part that should concern us all. Walk into almost any large enterprise and you'll find all these efforts being worked on right now – in at least five different systems, by five different teams, on five different definitions of what a "role" even is.

1) An “automation impact” tool handles the task and role automation assessment.

2) A job-architecture project – usually a stupidly expensive consulting engagement that lands every couple of years – handles the roles.

3) Excel and a bunch of analysts stuck in a dark basement room handles the strategic workforce planning forecast.

4) Talent Marketplace or L&D tools handle the reskilling and related development efforts.

5) An org-design tool handles the structure.

Five workstreams. Five taxonomies. Five reconciliation cycles.

But step back and look at what set all five in motion. It was one thing. A change in the underlying tasks. A single root signal, refracted through five systems that can't talk to each other, and the refraction itself, the endless translating of one system's "role" into another system's "role," is most of the effort and most of the cost.  

The majority of the team's intellectual capacity doesn't go into sharing strategic insights and guidance with the business and executive leadership, it gets spent on generating the preconditions for insights - the massaging and manipulating of data so that everything speaks at least a similar language.

So we don't actually have five problems. We have one problem, fragmented across a stack that was never built to treat it as one.

The technology to unify this exists now – we do it with Fortune 1000 clients every day. The harder problem is that most organizations are still organized to treat one signal as five.

We need to stop looking at this problem as a spear, and a rope, and a wall, and a tree etc... and instead see it as the single beast that it is. In our world, the pieces and parts are Automation Exposure, Work Architecture, Skills Intelligence, Talent Intelligence, SWP, and Org Design – all codified as separate solutions but all solving for the same root questions:

• What is the workforce I have?

• What is the workforce I need?

• How do I get from what I have to what I need?

• What scenarios should I plan for?

• What will that journey cost me?

And our version of the elephant? Workforce Intelligence. Every single solution mentioned above is some version of Workforce Intelligence - all trying to answer the same root questions from different attack vectors. What's required is a unified platform that combines these capabilities on top of a shared data architecture with shared models, methods, frameworks etc... That's what we've spent the last few years building, and it's what dozens of Fortune 1000 clients use now to solve these challenges.

The task became the unit of change last year. For most companies, the unit of response is still scattered across five systems that each define a role differently. We have a solution, but until we start conceptualizing the root issues and the ideal solution differently, no amount of tooling will help.

I'm curious whether you see it the same way. When the work changes at your company, does the response come out of one coherent foundation – or does it get assembled by hand across teams who rarely sit in the same room? What's actually worked for you?