The Assessment Is the Product: Organizational Consulting for the Lean, AI-Era Company

The Consultant the Lean Era Actually Needs

Something has changed about the shape of companies, and it changes what good consulting should look like. AI has let small teams do what used to require large ones, founders are deliberately keeping headcount lean, and the instinct to solve every problem by hiring or by retaining an outside firm indefinitely is fading. The companies I work with do not want a consultant who moves in. They want to understand their own organization clearly, fix what matters, and keep moving fast on minimum resources.

That makes the traditional consulting model a poor fit. The model built for an earlier era - a long, open-ended engagement where the consultant becomes a fixture and the monthly invoice becomes a habit - is exactly what a lean, fast company should not buy. What it should buy instead is clarity: a precise reading of where the organization is strong, where it is about to break, and what to do about each, delivered as a finished piece of work it can act on however it chooses.

That reading is the assessment. And the argument of this essay is that the assessment is not the appetizer before the real engagement. It is the product.

The Quiet Problem With the Retainer

I do not believe in the open-ended retainer, and the reason is structural rather than moral. A retainer pays a consultant to remain present, which is a different thing from paying them to solve the problem. Over time the incentive quietly bends toward continuation: the work expands to fill the engagement, dependence is rewarded rather than designed out, and the moment when the company could confidently run on its own keeps receding because no one is paid for it to arrive.

The better model, particularly now, is a clear diagnosis followed by well-scoped projects taken on as they are genuinely needed. Some of those projects the company runs itself. Some are worth embedding a senior operator for, on a defined piece of work or a fractional basis sized to the actual need. Some belong to a different specialist entirely. None of them require a meter running quietly in the background for months. This keeps the relationship honest, and - counterintuitively - it is what makes the relationship last: a consultant who earns the next project by making the last one pay off is far more valuable to a company than one whose value is assumed because the invoice keeps arriving.

So the question is not "how do we keep the consultant engaged?" It is "what is actually wrong, what will it cost us, and what is the smallest set of moves that fixes it?" That question is answered by an assessment, and the answer has value whether or not the consultant is ever hired again.

What a Diagnosis Prevents - the Money

The case for the assessment is easiest to make in the language of money, because the mistakes it prevents are not abstract. They are some of the most expensive things a company can do, and they are almost always done blind - which is to say, they are done because no one had a clear picture of the organization's real structure before acting.

Unnecessary layoffs the company later reverses

The clearest current example is the layoff made without understanding what the organization actually depends on. The recent wave of AI-driven cuts has produced a striking pattern: roughly 55% of employers report regretting layoffs they made for AI (Forrester), and a large share have quietly rehired - about 29% for the exact same roles they had just eliminated (Robert Half), with Gartner projecting that around half of companies citing AI for headcount reductions will rehire for similar functions by 2027. The most common stated reason is not that the technology failed; it is that the cut destroyed institutional knowledge the company could not see it was relying on.

This is expensive twice. Replacing an employee costs roughly 50% to 200% of their annual salary, according to Gallup and SHRM, and replacing a leader or manager sits near the top of that range. So a wrong cut is a six-figure error before the role is even refilled, and the knowledge that left often does not come back with the rehire. An assessment identifies which roles carry hidden structural load - the people decisions and processes silently depend on - before anyone is cut. When a reduction is genuinely necessary, it then targets the right places and does not have to be undone.

The sudden loss of a key person

Most companies have at least one person whose departure would stall the organization, and most have never named who that is. SHRM's own work on the loss of key employees calls this "the myth of replaceability": the assumption that anyone can be backfilled understates how much an organization runs on specific individuals and the relationships, judgment, and undocumented knowledge they carry. When that person leaves suddenly - and Gallup estimates voluntary turnover already costs U.S. businesses on the order of $1 trillion a year - the cost is not just the replacement. It is the projects that stall, the decisions no one else can make, and the customers who notice.

An assessment makes single-point-of-failure risk explicit while there is still time to act on it cheaply: by documenting what lives in that person's head, distributing decisions that have over-concentrated, and removing the dependency before it becomes a crisis rather than after.

Knowledge that lives only in people's heads

Closely related, and just as invisible, is knowledge risk: the critical know-how that exists as habit and memory rather than as anything written down or transferable. It is the reason the AI layoffs backfired - the institutional knowledge was never captured, so cutting the person erased it. It is also the reason onboarding is slow, decisions get re-litigated, and the same problems get re-solved from scratch. None of this shows on a dashboard, and all of it compounds as the company grows. The assessment surfaces where knowledge is dangerously concentrated and what it would take to move it into the system, where it stops being a liability and starts being an asset the whole organization can use.

Decision bottlenecks

A bottleneck is a point where everything waits - for one person's approval, for one overloaded team, for a decision loop that takes a week to close. It is one of the most common findings in a growing company and one of the least visible, because each individual delay looks reasonable and only the cumulative drag on the whole organization reveals the cost. Founders are frequently the bottleneck themselves, having never handed off decisions they made well when the company was small and cannot make fast enough now. An assessment locates the bottleneck precisely, which is the prerequisite for relieving it - because the instinct to "add more people" usually makes a bottleneck worse, not better.

Architectural and process risk

Finally, the structural layer: systems and a data foundation built for a tenth of the load the company now carries, and processes held together by tribal knowledge that work at the current size and will not survive the next one. These are the risks that feel fine right up until they do not - the architecture that suddenly cannot scale, the manual process that breaks the week volume doubles, the integration that fails under real load. They are also the cheapest of all to address early and among the most expensive to address in a crisis, because by then they are load-bearing. A diagnosis catches them while they are still a design decision rather than an emergency.

How I Run the Assessment

An assessment is only useful if it understands the specific company rather than scoring it against a generic checklist, so it starts by learning the organization itself: its culture, its environment, the sector it operates in, and - most importantly - the vision its leaders are building toward. I am not there to impose a playbook. I am there to understand what makes this company work and where its leadership wants to take it, and then to design the infrastructure that lets it get there with innovation, efficiency, productivity, and agility intact.

The lens is always people, processes, and technology together, because real organizational risk almost never lives in only one of them. A bottleneck that looks like a process problem is often a people-and-decision-rights problem. An architecture risk is often really a knowledge risk about who understands the system. Looking at the three in isolation is how most reviews miss the thing that actually matters. The work is interviews, a review of documents and systems, and observation - on-site or remote - of how the organization actually operates, not how its org chart says it does.

Fit, not overhaul

The part of the method companies notice most is what I do not do: I do not arrive and rebuild everything. I read how the organization already works - its real flows, the systems it already runs on, the habits that are quietly load-bearing - and I design solutions that fit that reality rather than replacing it. Most of what needs to change can be done with small, deliberate adjustments rather than a disruptive overhaul, because the existing flows usually already contain most of the answer; they need to be reinforced and pointed at where the company is going, not torn out.

The other half of the method is planning ahead. The whole point of catching what will break before it breaks is that the fix can be installed quietly, now, while it is still small - so that when the growth actually arrives it feels fluent instead of like a shock. Done well, the change is almost invisible from the inside, and from the outside it looks like the company had been preparing for that moment all along. That is the goal every time: growth that lands seamlessly, on infrastructure that was ready for it before anyone needed it to be.

Where the method comes from

This conviction is not theoretical. It started long before consulting - in streamlining events as a young organiser, where one person had to make a complicated thing run smoothly without a big team behind them - and it carried through years of building products and infrastructure from zero. The systems I built held as they scaled, and they held without large teams or heavy, complicated processes, because they were designed lean from the start and prepared for the next stage before it came. That is the same thing I now build into other people's organizations: structure that carries growth without bloat.

Calibrated to your sector

The right answer also depends heavily on what kind of company this is, and matching the solution to that context is part of the work, not an afterthought. A fintech operating in a highly regulated, competitive, and fragile environment - where almost everything has to be monitored, controlled, and approved - needs a very different design from a small startup whose entire advantage is the ability to move fast. The same finding can call for tight controls in one and deliberate looseness in the other. I have operated inside both kinds of environment, which is why the recommendation is shaped by the company's real constraints - culture, budget, market, regulatory reality, and the leaders' vision - rather than by a template. Preserving what already works is as much the job as fixing what does not; the aim is to let the company's own direction scale, never to replace it with someone else's idea of how a company should look.

What You Get at the End

The deliverable is not a generic report that restates what you already know in more words. It is a prioritized, specific picture of your organization that you can act on. Concretely, at the end of the assessment you have:

That last point is the one that matters most for trust. The plan is deliberately written to be executable without me. If the most valuable thing I can hand a company is clarity, then that clarity has to be usable on the company's own terms - otherwise it was never clarity, it was a dependency in disguise.

How to Act on the Recommendations

What happens after the assessment is a set of choices, made per recommendation rather than as a single package. There are three honest paths for any given item.

Internally. Many findings come with a fix clear enough that the company's own team can execute it with the plan in hand. This is not the consolation prize; for a great many recommendations it is the right answer, and the assessment is deliberately designed to make it possible. A company that can fix its own problems with a good map is in a stronger position than one that needs an outsider to hold the pen.

With me. For the items where embedded, senior execution genuinely adds value - a transition that needs running, a build that needs operator judgment, a function that needs a steady hand for a few months - I take it on as a defined project or a fractional engagement scoped to the actual work. Not a retainer. A piece of work with a beginning and an end.

With someone else. Where a specific specialist is the right fit - a security expert, a particular engineering skill, a domain I am not the best person for - the plan is written so any competent person can pick it up and run it. I would rather a company get the right help than keep the work inside an engagement where it does not belong.

Across all three, the sequence is the same discipline the assessment already established: address the irreversible, high-cost, soon-to-bite items first; prepare for the medium-term ones; monitor the rest. The plan tells you which is which, so the post-assessment period is not a fresh round of guessing - it is execution against a map you already trust.

Why the Assessment Builds a Long-Term Relationship

It might seem that designing the work to not lock a client in would shorten the relationship. In my experience it does the opposite. A company that has had one honest, useful diagnosis - one that paid for itself, respected its budget, and left it with something it owns - is a company that comes back when it actually needs the next thing, and refers others while it waits. Trust compounds faster than a retainer ever could.

So the assessment is the foundation of the relationship precisely because it asks for nothing beyond itself. It is the first piece of a long collaboration that proceeds project by project, as needs arise, with the company always in control of how much help it wants and from whom. That is a better deal for the company, and over a long enough horizon it is a better business for the consultant - because it is built on results that are easy to point to rather than on a presence that is hard to justify.

The Economics of Skipping It

The argument against doing an assessment is usually that the company is too busy, or that it already knows its problems. Both are versions of the same bet: that whatever is wrong can wait, and that acting on instinct will be good enough. The research on layoffs is the clearest evidence of how that bet pays out - a majority regretted, a large share reversed at full cost, institutional knowledge destroyed and not recovered. The same logic applies to every category above. A problem caught early is a design decision; the same problem caught late is a crisis, and crises cost an order of magnitude more to resolve because by then they are load-bearing and the cheap options are gone.

Set against that, a fixed-scope diagnosis is one of the lowest-risk investments a growing company can make. It has a defined cost, a defined timeline, and its entire purpose is to find the expensive problems while they are still cheap. The math is not subtle: preventing a single wrong layoff, or one avoidable key-person departure, returns the cost of the assessment many times over - and those are the routine findings, not the exceptional ones.

This Is the Work I Do

The Scale Readiness Assessment is the front door to everything I do, and for many companies it is the only thing they need from me. It runs over about six weeks, combines interviews, document and systems review, and on-site or remote observation, and ends with the prioritized plan described above. It is priced scoped to your budget scaling with team size, and it carries no obligation to engage me for anything afterward.

If something in this essay described your company - a key person you quietly depend on, a bottleneck you have stopped noticing, a system you suspect will not survive the next stage, or a layoff you are considering and are not certain about - that is exactly the kind of thing the assessment is built to make visible before it becomes expensive.

Related Reading

• The 10x Rule - Why every organizational problem you don't catch now costs roughly ten times more to fix at the next stage of scale. The economic argument underneath the assessment.
• The Fractional Future - Why a senior operator plus AI now beats a full-time hire for the work that follows an assessment, and how to size the commitment to the need.
• The Deliberate Slow - Which decisions must stay slow even in the AI era - directly relevant to the layoffs and architecture risks an assessment surfaces.
• Use Cases from the Field - Real patterns from organizational reviews across regulated and operational industries.

Sources

• Gallup, "This Fixable Problem Costs U.S. Businesses $1 Trillion" - voluntary turnover costs U.S. businesses roughly $1 trillion a year; replacing an employee costs about one-half to two times annual salary, with leaders and managers near the top of that range.
• Society for Human Resource Management (SHRM) - benchmarking on cost-per-hire (roughly $5,475 non-executive, $35,879 executive) and the commonly cited 50-200% of salary replacement cost; SHRM Executive Network, "The Myth of Replaceability: Preparing for the Loss of Key Employees."
• Forrester - industry research indicating roughly 55% of employers regret AI-driven layoffs.
• Robert Half and Careerminds - research indicating a large share of companies that cut roles for AI have rehired, including about 29% for the exact same roles; the leading stated reason is loss of institutional knowledge.
• Gartner - projection that about half of companies citing AI for headcount reductions will rehire for similar functions by 2027.
• Note: market-research and outplacement-firm figures on the AI-layoff reversal are recent (2024-2025) and directional; they are cited as industry estimates, not settled statistics.