Augment your people. Don’t try to replace them.

There are two things people mean by ‘augmentation’, and the AI moment has blurred the line between them.

The old meaning is staff augmentation: you bring in outside people - a contractor, a fractional operator, an agency - to add capability you don’t want to carry as permanent headcount. The newer meaning is human augmentation: you give the people you already have sharper tools, so each of them does more, and does it better. For most of a career those sat in separate conversations, and they don’t anymore. A leader now faces the same question either way: where do you add capability, and whether you add it as people or as tools.

The loud answer in the market is ‘replace them with AI’, and it is wrong in a way that gets expensive fast. The boring answer is the one that holds up: augment the people you already have, and spend the freed-up time on the work only they can do.

The replacement fantasy is a budgeting error

The pitch to a C-suite is seductive because it is phrased as a cost line. Headcount is expensive and it shows. A model looks cheap, and most of the time it doesn’t show at all. Swap one for the other and the spreadsheet looks better this quarter.

In practice you almost never replace a whole role. You replace a task or two inside it. A person in a revenue team does maybe a dozen distinct things in a week, and a model handles two or three of them well and is useless at the rest. Fire the person and you lose all twelve to save the two or three the model could have taken off their plate anyway. You also lose the thing that does not show up on the org chart: the judgment about which of the twelve mattered this week.

The firms that pull ahead over the next few years will be the ones who got more out of the people they kept. Cutting headcount hardest is not the same thing, and the difference only shows up if you invest in it on purpose.

What the research actually shows

The evidence is more interesting than either the hype or the panic. Two findings are worth a board’s attention.

The first comes from a large field study of customer-support agents given a generative AI assistant, run by researchers including Erik Brynjolfsson. The gains were real, but they landed unevenly. The biggest jump went to the newest and lowest-performing agents; the tool pulled the bottom of the distribution up toward the top and spread the know-how of the best people out to everyone else. The strongest performers barely moved. What augmentation mostly does is raise the floor.

The second is a Harvard and BCG study of management consultants using a frontier model, sometimes called the ‘jagged frontier’ work. On tasks that sat inside the model’s competence, augmented consultants were faster and produced better work. On tasks that sat just outside it, the ones who leaned on the model did worse than the ones who did not - because they had outsourced a judgment the model was not equipped to make, and did not notice. So the lesson isn’t really about whether to use AI. What matters is knowing which tasks sit inside the model’s competence and which sit just outside it, and that judgment is a human skill you have to keep paying for.

Put the two findings together and the strategy is fairly plain. Use models to take on the routine work and bring your newer people up faster, and protect the judgment that decides what counts as routine in the first place, because that is the part you can’t automate away.

What augmentation looks like in a revenue engine

The concrete version comes from the systems we actually build. The shape barely changes from one to the next. The model does the first draft or the first pass, and a person owns the decision and the outcome.

• Enrichment and research. A model assembles a company’s profile from scattered signals in minutes instead of an afternoon. A person decides whether the read is right and whether the account is worth the team’s time. (More on doing this without the guessing: CRM enrichment without the hallucinations.)
• Lead qualification. A model reads the record and writes a plain-English rationale for why a lead looks strong or weak. A rep reads the rationale and makes the call. A mystery score just gets ignored, whereas a written sentence can be checked and argued with. (How we build that.)
• Notes, summaries and drafts. A model turns a messy call into structured CRM fields and a first-draft follow-up. The person edits it into something true and sends it under their own name. That takes roughly 80% of the grind off the plate, and the bit that still needs a person is where the time now goes.

In every one of these the person still does the thinking. What drops away is the typing, which frees them up for the harder calls only they can make.

The models, and where the human stays

The tools are genuinely good at a narrow, valuable band of work: reading and summarising the text you hand them, and drafting from it. Anthropic’s Claude, OpenAI’s ChatGPT and Google’s Gemini are all, at heart, extraordinary readers and writers of text. That is most of what a revenue team does all day: taking one kind of text, a thread or a transcript, and turning it into another, a summary or a next step.

What they are not is accountable. A model can’t be the person a customer trusts, and it can’t carry the context nobody ever wrote down. It won’t own the consequence of being wrong, and it has no way to decide what matters when two goals pull against each other. A bigger model doesn’t close those gaps, because they are a different kind of work altogether. An augmented team keeps all of it on the human side of the line.

Why AI cannot replace your people - the reasoning, not the slogan

‘AI won’t replace you’ usually gets said as reassurance. It works better as an argument, because the argument is the thing that tells you where to put your money.

Judgment under ambiguity. Most real decisions are made with missing information and competing goals. A model can tell you what is typical. It can’t tell you what is right for one particular customer in one particular quarter, because ‘right’ here depends on things that were never in the training data and never will be.

Accountability. Someone has to be answerable when it goes wrong. You can’t fire a model or promote it, and you can’t stand it in front of a board to explain itself and mean it. In the end accountability is a person choosing to stand behind an outcome.

Unwritten context. The most valuable thing your best operator knows is not in any document. It is a feel for which deals are real and which numbers are lies, plus the instinct for which colleague to call when something looks off. That knowledge gets built by doing the work, and it fades the moment you stop letting people do it.

Relationships and taste. Buyers buy from people. Judgment about what is good - a good message, a good hire - is a human faculty, and it only holds up if people keep exercising it. Outsource it wholesale and you don’t save money. You lose, a bit at a time, the ability to tell good from bad.

This is the same line we draw in our AI manifesto. Use AI where it genuinely beats a hand-written rule, and keep a person on any decision that carries a real consequence. Nothing goes out that a human hasn’t read.

What a leader should actually do

Invest in both, on purpose. Give your people the tools that take the routine off their plate, and put the time you free up back into the human work, the judgment and relationship-building that don’t automate, rather than pocketing it as a headcount cut. Level up your newer people deliberately, because that is where the research says the gains are largest. And draw the line, out loud, on what stays human, so nobody has to guess whether the model is allowed to make a decision it should not.

Staff augmentation and human augmentation were always the same instinct: add capability where you need it, without pretending the people are the problem. AI doesn’t change that instinct, it just raises the cost of getting it wrong. So bet on your people and hand them better tools. The firms that do will pull ahead of the ones still trying to subtract their way to a better number.

Common questions