Every transformation has a moment where someone falls in love with the tool. The demo is clean. The interface is quiet and confident. Someone in the room says the word intelligent, and for a few seconds the future looks like a screen. This is the moment the program is most likely to go wrong, and it goes wrong so pleasantly that no one notices.
Because a transformation is not a thing you acquire. It is a thing you decide to become. And nothing on the screen can make that decision for you.
The industry gives itself away in how it names things. The HRIS program. The migration. The platform rollout. Named, in other words, after the object being installed rather than the capability being built. A program named after its object will, quite naturally, measure itself on whether the object arrived: on time, on budget, live. Whether the organization became anything is a question the name was never designed to ask. And so the program is declared a success at precisely the moment its real work has not yet begun, and everyone moves on before anyone can notice that nothing underneath has changed.
The seduction of the tool
It is worth being honest about why the tool seduces. It is legible. You can see it, price it, schedule it, and point to it when it is done. A capability is none of those things. You cannot point to "we are now an organization that hires the right people well." You can only point to the applicant tracking system you bought on the way to not quite becoming that.
So the tool becomes a proxy for the outcome, because the tool is the part you can hold. And the coolness of it, the genuine and not-imaginary power of what the new technology can do, makes the proxy feel like the real thing. The danger is not that the technology is oversold. Often it is exactly as good as promised. The danger is that its goodness answers a question you were never supposed to ask first.
Start from the end
The only serious first question of any transformation has nothing to do with technology. It is this: when this is finished, what must the business be able to do that it cannot do today, in order to meet what its customers actually need? Answer that, precisely, and the design has a spine. Skip it, and you have bought a very expensive way to keep doing what you already did, only faster and with a support contract.
This is why the choice of tool is not the first choice. It is close to the last. What you want to become sits upstream of everything: upstream of the platform, the vendor, the module, the button. The technology is downstream of the decision, and a great many programs run in the wrong direction, letting the tool they happened to buy define the future they are allowed to want.
Notice that this holds regardless of the function. In HR the thing you are trying to become might be an organization that hires well, or develops people rather than merely tracking them, or pays people in a way it can actually stand behind. In supply chain it is sensing demand before it arrives, or moving goods with less friction, or holding less stock without stocking out. In IT it is resilience, or speed, or a lower cost of simply staying alive. The function changes. The tools change. The first question does not move. What must we be able to do, and to be, when this is done? A transformation that cannot answer that in a sentence has not started designing. It has started shopping.
Two kinds of knowing
Say the plainest true thing about the technology available now: it is a genuine accelerant. Used well, it does not improve a process by a margin. It changes what is possible, and it does so at a speed that would have looked like fantasy only a few years ago. That is real, and it is worth wanting. Nothing in this argument is an argument against the tool. It is an argument about the order of operations.
Because an accelerant needs two kinds of knowing to return its value, and organizations reliably acquire only the first.
The first knowing is operational. What do I do with this? Which inputs, which steps, which buttons, in which order. And, the part almost everyone forgets, what do I stop doing, because the old motions are now waste and will quietly re-import the old process into the new machine if no one names them. This first knowing is teachable, and vendors teach it well, because it is the knowing that sells and renews licenses.
The second knowing is harder, and no one sells it, because there is no product to attach it to: what has to be true for the technology to work at all? What conditions must already exist in the organization for the tool to work rather than sit there, capable and idle and resented? Data clean enough to trust. A decision a human used to make and now has to be genuinely willing to hand over, not merely permitted to override. A process that survived for years on ambiguity and now has to be made explicit before a machine can touch it. A manager who understands that the point was never the dashboard.
Plant a capable tool in an organization that has not prepared any of that ground, and it will fail in the soil, no matter how good the tool looked on the shelf. This is the failure that does not announce itself. The system is live. The project is closed. And value simply never arrives, quarter after quarter, because the conditions it needed were never part of the plan. The plan was about the tool. The value was always about the ground.
What the tool cannot decide: a hiring journey
To see how much of a transformation lives in decisions the software will never make for you, take a single journey and look closely: the path an organization travels from the reputation it holds in the market to a person who says yes and starts work.
Notice first how many actors are in it. The employer's reputation and culture. The candidate. The sources where the right talent actually is. The channels that carry the message. The hiring manager. The recruiter. The compensation framework. The future coworkers. The candidates who succeed and the ones who are turned away. The administration that quietly processes all of it. That is not a funnel. It is a small society, and each member of it meets the organization at some point and comes away with a judgment about what kind of place it is.
Each of those meetings is a moment of truth. And each moment of truth is really a decision, turning on factors that refuse to hold still: the company's own culture, the culture of the place where the role sits, the seniority and shape of the job, the scarcity of the people who can do it. A rejection written one way in one country reads as respect, and the same words in another read as insult. The same moment is not the same moment twice, and a system configured once, centrally, does not know the difference unless someone taught it to.
Now lay the technology across that map, and the argument of this whole piece becomes something you can see. At every one of those moments there is a possible data point that could inform a better decision. And at every one there is a prior choice that the software will not make for you: does a human create the most value here, or does the machine? Where does automation free a person to do the thing only a person can do, and where does it strip out the human presence that was the entire reason the moment mattered in the first place?
The tool does not tell you. It cannot. It will faithfully execute the choices you have already made about what your hiring should feel like and who it is for, and wherever you have made no choice, it will impose a default that someone else designed for someone else's company. Start the program with the platform, and you inherit those defaults without ever knowing you did. Start it with the journey, with the moments and what you have decided each one should be, and the platform finally becomes what it was always meant to be: an instrument that serves a design, rather than a design bent quietly to serve an instrument.
This is only one journey. Every function has its own, with its own society of actors and its own moments of truth. The lesson does not belong to recruiting. It belongs to the method. The value was never in the map alone. It was in the choices you made at each point on it, and in knowing which of those choices a machine should hold and which a person must.
The choice is the deliverable
So the real work of a transformation is not the installation. It is the sequence of choices about what the organization wants each of those moments to become, and then the harder choice of where a human and where a machine will best deliver it. That work is not glamorous. It does not demo well. There is no screen to show for it at the steering meeting. It is also the only part of the whole endeavor that no vendor can do for you, and therefore the only part that is genuinely, permanently yours.
This is what it means to say it starts and ends with you.
It starts with you because the first question, what do we want to become, is one that only the organization can answer, and answering it is not a technical act. No amount of platform will produce it, and no consultant can hand it over, though many will offer to.
And it ends with you because when the program is finished, when the consultants have gone and the vendor has moved on to the next account, what remains in the building is one of two things. Either a capability you now hold and can use long after anyone remembers which software delivered it, or an expensive system you never quite grew into, humming away, technically live, quietly returning nothing. The technology withdraws either way. What it was planted in is what stays.
So the question to carry into the next steering meeting is not which platform, and not how soon. It is the one that is easy to skip precisely because no vendor is in the room to answer it: if this transformation succeeds completely, what will we be able to do that we cannot do today, and have we designed a single moment of it around that answer, or only around the tool?
Wherever the honest answer is "only around the tool," the transformation has not started yet. It has only been purchased.
Clover & Myrtle is a vendor-independent HR transformation practice. The work the industry forgets: before the tool, and underneath the deadline.
cloverandmyrtle.com/journal