It is the Thursday before the quarter closes, and the deal is going to land. Everyone on the forecast call knows it. The rep has walked us through it three weeks running, and tonight he has the easy calm of a seller sitting on a verbal yes.
"It's done," he says. "Champion's all in, he's been selling it internally for us. I met the economic buyer in person last week; she gets the business case. Legal's nearly through redlines. Worst case, it's Wednesday."
Someone asks about the competitor. "Dead. They were never really in it." Someone asks about budget. "Approved in the last planning cycle." The deal is Commit. It has been Commit for a month. We move on.
It does not close on Wednesday. It does not close that quarter. Six weeks later it is "pushed," then "re-engaging in H2," then quietly gone, a reorganisation, a new VP with her own roadmap, a budget swept into a different priority. In the post-mortem the rep is truly baffled, and if they are honest, so is everyone else on that call. Because by the book, by every letter of the framework we all swear by, that deal was as qualified as a deal gets.
I have sat on both ends of that call more times than I would like to admit. Sometimes I was the leader nodding along. Earlier in my career, I was the rep. And after enough of them, you stop reaching for the easy explanation, the rep got happy ears, the champion went dark, and you start asking a more uncomfortable question. What if the deal was never as healthy as it looked? What if the instrument we used to measure it was simply pointed at the wrong thing?
The framework that taught us to ask
Let me be clear about what I am not doing here. I am not here to bury MEDDIC.
If you have sold enterprise software in the last thirty years, you have lived inside it. Metrics, Economic buyer, Decision criteria, Decision process, Identify the pain, Champion, extended later, as MEDDPICC, with the Paper process and the Competition. It was forged at PTC in the early 1990s and then carried, through PTC and BladeLogic and BMC and a long list of companies after them, into the qualification discipline that now underwrites most of the enterprise playbook. John McMahon, more than anyone, is the reason it spread; if you have not read The Qualified Sales Leader, read it, and notice that he does not open with a framework, he opens with a scene, a leader walking into a mess. He earned the right to teach it the hard way, and a generation of us learned it from him and the people around him. I count myself part of that community of practice, and many of its staunchest advocates among my closest friends. So take what follows as a colleague arguing for an upgrade, not an outsider throwing rocks.
And it deserves the respect, because underneath the acronym MEDDIC is a genuine piece of research. Its creators did something quietly scientific: they looked across a large number of won and lost deals and asked what the winners had in common. The answer turned out to be structural. Winners had a quantified business case. Winners had reached the person who really controlled the money. Winners understood how the decision would really be made and what paper it had to clear. Winners had an insider who argued their case when they were not in the room. That is pattern recognition of a high order. It took the chaos of complex sales and distilled it into a handful of things that truly move the outcome, and it gave a whole profession a shared language for the only question that matters in a pipeline review: is this deal real?
The instinct was right. It is the instrument that has not kept up.
The flaw is not the model. It is the sensor.
Here is the part where the framework's advocates rightly bristle, so let me say it carefully. Every letter in MEDDPICC is, mechanically, a claim made by a human being, usually the human whose number and forecast depend on the answer being favourable. Is there an economic buyer? The rep says yes. Is the champion real? The rep says yes. The framework is, at bottom, a self-report survey, and it was designed that way for an honest reason: when MEDDIC was created, the only data about a deal was whatever the seller typed into the CRM. There was no other instrument in the building. So the framework encoded abstractions, a person's assertions about reality, because assertions were all anyone could capture.
This is not a story about reps lying. It is a story about something far more stubborn: the limits of human perception, and they have been measured for half a century. In 1974, in Science, Amos Tversky and Daniel Kahneman published the paper that founded modern behavioural economics, and its finding was not that people are stupid but that they are predictably wrong, that under uncertainty the mind quietly swaps hard questions for easy ones and reaches confident answers through shortcuts it cannot feel itself taking. Kahneman spent the next forty years cataloguing the ways, and the one that should keep every sales leader awake he abbreviated to five words: what you see is all there is. The mind builds a coherent, confident story out of whatever information is in front of it, and is structurally blind to the information that is missing. Worse, the more coherent the story, the more certain it feels.
Read the forecast call again with that in mind. A strong last conversation, a warm champion, a business case the rep believes in: that is more than enough raw material for the mind to assemble a complete, convincing, green-across-the-board story, and to feel sure of it. Not because the rep is careless. Because that is how the instrument works. And the second, slower, supposedly rational part of the mind, the part that is meant to check the story, mostly just rubber-stamps it. We do not verify our way out of a conviction we have already formed. We go looking for the evidence that confirms it.
The signals we trust are the ones most likely to fool us
It gets harder. Even a disciplined seller who interrogates every letter is reading human signals, what the champion says, how the room felt, and those are exactly the signals a buyer is most able to distort, usually without meaning to.
Consider the champion who tells you everything is on track. Perhaps they are managing your feelings; few people enjoy delivering bad news to someone they like. Or perhaps, and this is the dangerous case, they sincerely believe it, because they are sitting inside a pocket of the organisation that has quietly talked itself into a shared, over-confident view. Irving Janis gave that its name in 1972: groupthink, the tendency of a cohesive group under pressure to converge on a consensus and suppress its own doubts, complete with an illusion of unanimity that no one inside the room can see. Your champion may be giving you a perfectly honest report of a consensus that is itself wrong. You can interrogate a sincere champion flawlessly and walk away with a confident false positive. The model was sound. The instrument was a human being, doing their best, inside a system designed to mislead them.
The forecast inherits the optimism
Now layer on what we know about how people predict the future, because a pipeline is nothing but a stack of predictions.
Kahneman called it the planning fallacy: our forecasts cluster around the best case and ignore what usually happens. His favourite illustration is almost too on the nose for sales, American homeowners in 2002 expected a kitchen remodel to cost around eighteen and a half thousand dollars; the average bill came in at nearly thirty-nine thousand, more than double. They were not lying to themselves on purpose. They were doing what we all do: taking the inside view, this project, this story, and ignoring the outside view, the base rate of what happens to projects like this one. A rep's forecast is the purest inside view there is. The antidote has always been the outside view: what tends to happen to deals that look like this. That is precisely the thing no single seller can hold in their head, and precisely the thing data across hundreds of deals can.
Then there is the cruelty of memory. Nassim Taleb calls it silent evidence, we study the survivors and never count the quiet failures, so we systematically overrate our read of what causes success. Every "the champion was strong, that's why we won" pattern in your head is built only on the deals you remember winning; the lost deals that looked identical never made it into the model. And once a rep has poured months into an opportunity, a different bias takes the wheel. Kahneman observed that a team invested in a direction for months grows progressively less open to evidence that contradicts it. The deal stops being a hypothesis to test and becomes a position to defend. The rep has, emotionally, already banked it.
When it finally collapses, we do the most human thing of all: we build a tidy story about why. Taleb calls this the narrative fallacy, and his sharpest example is the morning of 11 September 2001, within hours, commentators were confidently explaining why the attack had been inevitable, though not one of them had predicted it the day before. Our post-mortems run the same way. "We lost because procurement got involved late." Perhaps. Or perhaps the deal had been quietly cooling for weeks and procurement was simply where the cooling finally became undeniable. The story arrives after the fact and gives us the comforting illusion that we understood the thing all along.
What an enterprise deal really is
Step back far enough and you see that we have been describing the wrong object. A complex B2B sale is not really a checklist being completed. It is a belief spreading through a group of people.
George Akerlof and Robert Shiller, one a Nobel laureate, the other the man who called two of the great asset bubbles, wrote a whole book arguing that economies do not run on cold calculation but on what John Maynard Keynes called animal spirits: confidence, fairness, corruption, money illusion, and, above all, the stories people tell each other. Their central mechanism is the confidence multiplier. Confidence, they point out, is not a forecast; it is a form of trust, credo, the Latin for I believe, and it feeds back on itself, so small shifts in sentiment produce outsized swings in behaviour. They also note, almost in passing, something every enterprise seller has felt: people feel real pressure to adopt the views of the groups they belong to.
That is a description of a deal. Winning one is the engineering of a contagion, a positive belief about you that has to spread across a tightly coupled cluster of professionals and reach critical mass before a signature is possible. Which means a deal is vulnerable to exactly what every contagion is vulnerable to: a change in the surrounding conditions that slows or reverses the spread. The reorg. The new VP. The competing priority. The belief stops propagating, and then it starts to recede.
Contagions obey laws, and the laws are measurable
Here is where it stops being a metaphor and becomes a science, and it is the part I would stake the whole argument on.
For two decades, Nicholas Christakis and James Fowler have studied how things move through human networks, most famously by mining the Framingham Heart Study, 4,739 people, more than fifty thousand social ties, tracked over twenty years. What they found is that behaviours and feelings spread between people like infections, and they spread according to a rule they named the three degrees of influence. Your state ripples to your friends, to their friends, and to their friends, and then it fades. When one person in the network became obese, a friend's odds of following rose by about 57 percent; a friend of that friend, by around 20 percent; the third degree out, by roughly 10 percent. Past three hops, the signal dies. Happiness behaved the same way, each happy connection measurably lifted your own odds of being happy, an effect that reached three degrees and lingered for up to a year.
And the reason it decays past three degrees is the detail that matters most for sellers. Christakis and Fowler put it down, in part, to a kind of social friction, the signal corrupts a little each time it passes from person to person, like a game of telephone. Sit with that. Your champion, three hops from the executive who will really decide, is not a reliable narrator of that executive's intent. They cannot be. The signal has degraded by the time it reaches them, and degraded again by the time it reaches you. You are not being lied to. You are reading a photocopy of a photocopy.
The sociologist Mark Granovetter adds two more pieces that every complex-deal seller will recognise the moment they are named. In his famous work on the strength of weak ties, he showed that the information that truly changes your situation usually arrives not through your closest contacts but through your peripheral ones, which is why the stray comment from a stakeholder you barely engaged is so often the thing that turns a deal. And in his threshold model of collective behaviour, he showed that group adoption is not gradual; it is a cascade. Nothing happens, nothing happens, nothing happens, and then, once enough people have quietly moved, everyone moves at once. It is exactly why a deal can look flat and stuck right up until the week it suddenly closes, or the week it suddenly dies.
Why the data sees the fever break first
Which brings us back to that Thursday call, and to a turkey.
Taleb's turkey is fed by the farmer every single day, and every day its confidence in the farmer's benevolence grows, supported by an ever-longer run of evidence, until the Wednesday before Thanksgiving, when its confidence is at its absolute peak and its prospects are at their absolute worst. The deal that "never looked healthier" the week before it died is that turkey. Confidence and danger peak together, because the confidence multiplier runs in reverse just as violently as it runs forward; belief does not ebb gently, it breaks.
And here is the hard truth about relationships in sales. The rep with the closest relationship in the account is structurally the last person able to call the turn, because accepting that the contagion is slowing means accepting the loss of the deal they have already banked, and we have sixty years of behavioural research telling us how badly the human mind handles that. But the slowing leaves fingerprints long before anyone will say it out loud. Replies that used to come in an hour now take three days. Threads that used to widen, more names, more forwards, start to narrow. Stakeholders who were active go quiet. The fever is breaking in the data weeks before the best-connected human in the deal is willing to admit it. Early warning is the one thing a close relationship is worst at providing, and the one thing behavioural data is best at. It is no accident that the modern forecasting stack, Clari and the tools around it, now pulls interaction data straight into the forecast. They worked out, empirically, that what people do predicts revenue better than what reps say.
From self-report to observed signal
This is the whole move, and it is smaller and more respectful than it sounds. We are not throwing MEDDPICC away. We are keeping its hard-won structure and swapping its unreliable sensor, human assertion, for a reliable one: observed behaviour. Each letter survives; it just gets a better instrument.
Champion stops being a name in a field and becomes propagation, is belief spreading across the buying group without your having to push it? Economic buyer stops being a box the rep ticks and becomes verified senior engagement, is there evidence the person with the money really reviewed the business case before the evaluation closed? Identify the pain becomes the pattern of how a buyer behaved before the formal process even began, which is one of the most predictive and most ignored signals there is. Metrics and decision criteria, in any evaluation-led sale, become demonstrated usage, what the buyer's people really did with the product, not what they said they thought of it.
I have done this exercise on real pipelines. In one business I helped rebuild, we went back through years of won and lost opportunities and let the data tell us what had separated them, not what the deal reviews had claimed. The honest signals were almost never the ones the framework foregrounded. Deals where the economic buyer had truly engaged with the business case before the evaluation closed won far more often and closed materially faster. Deals that stayed single-threaded slipped. Early, self-initiated engagement before the cycle even opened was a quiet, powerful predictor of conversion. None of it contradicted MEDDPICC. All of it made MEDDPICC measurable.
A qualification model that reads the world
Push the idea to its conclusion and qualification stops being a survey the seller fills in and becomes a model that watches the deal. The scoring framework I have been building works across four layers of externally observable signal rather than rep narration, the organisational (growth, complexity, regulatory exposure, executive and data maturity), the market (trend alignment, competitive pressure), the behavioural (content and engagement patterns across the buying group) and the transactional (prior purchases, real responsiveness). A language model scores each opportunity against defined criteria; a knowledge graph surfaces the relationships a flat CRM cannot see; and a human feedback loop keeps the whole thing honest.
The philosophical difference from MEDDPICC is the entire point. MEDDPICC asks the seller what they believe. This reads what the world is doing.
What interaction data still cannot do
I want to be as clear about the limits as I have been about the flaws, because the people I learned this craft from will hold me to it.
Interaction data is not judgement. Whether a pain is worth solving, whether a champion truly has the political capital they appear to, whether to walk away from a deal that looks alive but smells wrong, those still demand a human who has sat across the table and can read a room. The model narrows where to look and tells you when to look harder; it does not do the looking. And the signals are motion-specific: what predicts a win in a usage-led platform sale is not what predicts one in a procurement-led enterprise deal, so they have to be learned for each motion and re-validated as buyers and markets drift, which, not incidentally, is exactly why this approach gets sharper the more motions you run it across. This is an upgrade to the instrument, not a replacement for the practitioner. It makes a good seller better aimed. It cannot make judgement unnecessary, and anyone selling you that is selling you snake oil.
Where this goes
MEDDIC was the right answer to the right question, what makes a deal real?, asked at a moment when the only available evidence was the seller's word. The question has not changed. The evidence has. We can now watch the contagion spread or stall in something close to real time, read the social friction in a thread, see the fever break before the most optimistic person in the deal will admit it is breaking.
So the next standard will not be a tidier acronym recited in a Thursday pipeline review. It will be a model that observes the deal and tells you, from behaviour rather than belief, whether it is what the rep hopes it is. The teams that win the next decade will forecast on what buyers do, not on what sellers hope.
Go back to that deal that died, the one that was green across every letter, the one nobody saw slipping. It was slipping, and it had been for weeks. The champion's replies had been getting slower. The new VP's name had started appearing on calendar invites the rep never mentioned because they did not seem important. The thread had stopped widening. Every fingerprint was there, in the data, while we sat on a call and told each other a confident story.
MEDDPICC taught us what to look for. Data, at last, lets us see it.
And yet everything I have described so far is the shallow end.
Interaction data, who replied, who attended, who opened the business case, is only the most obvious exhaust a deal gives off. It sits closest to the surface, which is why the forecasting tools reached for it first. But it is a sliver of the digital footprint a buying organisation now leaves, and a smaller sliver still of what a serious analytical discipline can do once you stop treating a CRM as the edge of the known world.
The footprint is far larger than the inbox
Think about everything that is knowable about a buyer that has nothing to do with whether they answered your last email. The shape and velocity of their hiring. The executives arriving and leaving, and the mandates they carry in and out with them. The technologies appearing in their stack. The regulatory weather forming over their industry. The financial pressure visible in their disclosures. The competitive moves bearing down on them. The way their own market is starting to talk about the problem you solve. Their product telemetry, where you have it, in a depth that goes far past "they logged in." Every one of these is a signal, and most sales organisations use almost none of them, or use them as the occasional hand-assembled "account plan" slide, which is to applied data science what a horoscope is to astronomy.
Because the larger gap is not data, it is discipline. Most of what passes for "AI in sales" today is a thin cosmetic skin, a bot that drafts an email, a model that colours a pipeline stage red. That is not what quantitative teams in other industries mean by working with data. They mean base rates and propensity models; survival analysis on how deals of a given shape really progress; anomaly detection that fires when an account breaks from its own pattern; causal inference, where the data allows, to separate the signals that cause wins from the ones that merely ride alongside them; credible intervals in place of a single hopeful number. Done properly, this is closer to what a trading desk does with market data than to what a CRM does with a deal record. We are at the very start of pointing that machinery at the revenue problem, and the gap between a team that does and a team that does not will not be a few points of win rate. It will be the gap between seeing the board and guessing at it.
So when I say interaction data has limits, I mean it the way an astronomer means the naked eye has limits. True, and beside the point, once you have built the telescope.
When the work starts making itself
Here is the shift that changes the practitioner's life, and it is already under way.
For the entire history of this profession the seller has been the factory. The tailored deck, the business case, the follow-up that picks up the third point the CFO raised, the mutual action plan, the QBR pack, the renewal narrative, all of it produced by hand, by the same person we also ask to build relationships, exercise judgement and carry a number. It is the least leveraged use of a talented human being ever devised, and it is the first thing to go.
Generative agents produce that material now, at almost no marginal cost and at a quality a good rep would have spent a day reaching. That inverts the job. The seller stops being the author of the collateral and becomes its critic, judgement applied at the point of leverage, reviewing and directing and approving, rather than at the point of production. Anyone who has managed people already knows the motion: it is the difference between writing the proposal yourself and editing a sharp junior's draft. You can do the second for ten deals in the time the first cost you for one.
Now feed that same practitioner the output of the analytical machine I just described, not a dashboard to go away and interpret, but the conclusion, with its evidence and its recommended move. This account's engagement has broken from its own baseline; three of your five contacts have gone quiet; here is the intervention, and here is who to call today. The rep's scarce, expensive attention is spent only where it is irreplaceable: the judgement calls, the rooms, the relationships, the moments that truly turn a deal. Everything else is produced and surfaced for them.
The consequence is not a productivity nudge. It is an order-of-magnitude change in how many opportunities one competent human can carry without dropping the ones that matter, and therefore in how large a team a given amount of revenue requires. The arithmetic of the sales organisation, more or less unchanged for decades, comes apart. Fewer people, each spanning a far larger portfolio, each amplified rather than buried.
A different job, and a deeper dependency
Two forces then meet, and between them they reshape the role.
The first is the slow death of the one-off close. As software moves to consumption and usage-based pricing, the money is no longer made at signature; it is made in adoption, expansion and retention, and commissions are following the money. The rep's job is less and less "win the deal" and more and more "create and hold the conditions under which value compounds inside the account": land, expand, renew, expand again. Which is to say the job becomes, even more nakedly than before, the management of a contagion, not a single belief spread once to get a signature, but a belief planted, spread, defended and re-spread across a living organisation, indefinitely.
And there is the irony at the centre of all this. Strip away the production work, point the data science at the deal, automate the touches, and what is left for the human concentrates onto the one thing this whole essay has been circling: the deliberate creation and amplification of positive belief across a coupled cluster of people. The instruments do not replace the contagion engineer. They unbury them.
Swarms, and the scale of belief
Which raises the obvious question. If a deal is the spread of a belief through a network, and we can now both see that network and produce tailored material against it at almost no cost, what happens when you stop doing the spreading one manual touch at a time?
This is where directed agent swarms come in, and it is the part with no real precedent. Historically a single human could nurture a handful of relationships in an account and hope the belief propagated through the rest of the cluster on its own, the "social friction" of Christakis and Fowler quietly degrading the signal at every hop the rep could not personally reach. A coordinated swarm of agents, directed by that human, can work the whole cluster at once: a different, individually relevant thread to each stakeholder, timed to each one's behaviour, sustained across the many months a complex deal breathes, every touch informed by what the data says about where the contagion is strong and where it is cooling. This is not spam at scale, it is the opposite. It is personalisation and persistence at a density no human team could hold, aimed by a human who sets the intent and keeps the judgement.
Let me be careful here, because this is exactly the territory where hype curdles into nonsense and does real damage. An ungoverned swarm turned loose on a buying group is not a revenue engine; it is a reputational accident waiting to happen, and the evidence is already in, the firms bleeding the most from AI right now are the ones who pointed it at unvalidated data and unsupervised decisions and walked away. A swarm is only ever as good as the intent that directs it and the data that aims it. Which is the entire point: the human's role does not shrink in this picture, it concentrates into the two things that were always the real job, judgement about what is true, and judgement about what is right. The machine carries the contagion at a scale that was, a few years ago, simply inconceivable. The human decides where it should go, and where it must not.
The job that is left
So here is the shape of it. The deal was always a contagion of belief moving through a tightly coupled group of people, MEDDPICC understood that in its bones, even when all it could do was ask a tired seller to describe the spread from memory. For thirty years we managed that contagion blind and by hand, one relationship at a time, and called the guesswork a forecast.
We are no longer blind, and it is no longer by hand. We can watch the belief move, through interaction data first, then through the far richer footprint and the real analytical depth most of the industry has not yet reached for. We can produce the material that feeds the belief without burning a person to do it. And we can spread and sustain that belief, through directed swarms, at a consistency that turns what used to be a heroic individual effort into something closer to engineering.
What is left for the practitioner is not less. It is the distilled core of what selling always was beneath the paperwork: the judgement to know which belief is true and worth spreading, and the human presence to carry it into the rooms where it matters most. Fewer of them, each carrying more than we could have imagined a decade ago, each more dependent than ever on the one thing no model will do for them, the manufacture of conviction in another human being.
If MEDDPICC taught us what to look for, and data has taught us to see it, the work now is to act on it at the scale the thing always demanded, and to hold on to enough judgement, in a world of machines that will do whatever we point them at, to be sure it is worth doing at all.
There is one more piece, and the industry will get it wrong first, because it is an organisational problem wearing a technological disguise, and organisations defend their shape harder than they defend their results.
Borrow a word from the data people: ownership
In data governance there is a role with a deceptively quiet name: the Data Owner. Not the person who stores the data or tends the pipeline, but the person accountable for a domain of it, its definitions, its quality, its fitness for the decisions that lean on it. The owner is the one who answers for whether the thing can be trusted.
The account executive of the next decade is the Data Owner of their territory.
That is a heavier claim than it first sounds. It means the seller stops being a passive consumer of whatever dataset the centre deigns to hand down and becomes accountable for the inputs themselves, every attribute, every source, every signal relevant to their patch. The freshness of the org chart. The accuracy of the stakeholder map. The product-usage feed, the intent sources, the technographic picture. Not because they enjoy data hygiene, but because it is their patch, their number, their contagion to read, and the person with the most context and the most at stake is the right one to own the integrity of the machinery, not the person sitting furthest from the deal. We have spent two decades treating data quality as something that happens to the frontline, done indifferently by someone else, in a system the frontline neither designed nor trusts. Ownership flips that. The signal becomes the seller's responsibility, because the seller is the one it fails.
Decentralise the intelligence, not just the data
The instinct, when a company decides to get serious about data, is to centralise. A data team. A centre of excellence. A master-data programme to make the definitions consistent. A platform, procured centrally, and a set of dashboards designed centrally and pushed down to the field. Some of that is necessary, shared definitions are the grammar everything else depends on, and master data management earns its keep. But as an operating model for intelligence, it is a bottleneck dressed as rigour, and I have watched exactly how it breaks.
At DataRobot we automated the building of machine-learning models, and the most instructive thing about it was the anticlimax. Automated model-building, run across enough problems, lands again and again on the same workhorses, a gradient-boosted ensemble, a random forest, because for most business questions that is simply what works. The mathematics was never the hard part. The hard part was the human queue in front of it. And the deeper truth, the one that took longest to admit, was that almost no one in the business wanted a model at all. They wanted the answer: which accounts to work and which to drop, which deal was cooling and why. Every time a central team of specialists sat between the question and the answer, you got latency, a backlog, and a quiet politics of whose request mattered most.
Now collapse that distance. Let the account executive put the question to an agent in plain language, in the tool they already live in, Slack, Teams, a WhatsApp thread, and have the agent do the procuring, the joining, the modelling and the explaining, then hand back not a model but an answer in a sentence. The bottleneck does not get optimised; it disappears. Voice agents take it a step further again, until interrogating the intelligence of your own territory is no harder than asking a colleague over your shoulder. None of this is speculative, every piece of it can be built today. What it takes is not a breakthrough but a redirection: moving the attention and the budget away from the next centrally procured dashboard and towards intelligence and workflow sourced to the frontline. Attention, it turns out, is all you need, in rather more ways than the field that made this possible intended.
This is mission command, not a new dashboard
If that sounds radical, it is only because sales has not yet caught up to a pattern the rest of the world settled decades ago.
The clearest statement of it is military. The doctrine of mission command, Auftragstaktik, in the tradition that formalised it, holds that headquarters should set the intent and the objective and then push the decision about how down to the officer at the front, because the front has the freshest information and the slowest, most fatal thing in a fast-moving conflict is a decision that has to travel back to a map room to be made. It worked well enough that the rest of the world spent the following half-century studying and copying it, and it is now the backbone of how adaptive organisations everywhere are taught to operate. Agile software development is the same idea in civilian clothes: stop trying to plan the whole thing centrally and up front; push decisions to small empowered teams sitting close to the work, because software is a complex, shifting domain where the centre cannot know enough in time. So is every other truly adaptive, complex-responsive process. They all converge on the same answer, when the environment changes faster than a centre can sense and respond, you move the sensing and the deciding to the edge, and you bind it together with clear intent rather than detailed control.
Enterprise revenue is exactly that kind of environment. Buyers move, stakeholders churn, markets turn, the contagion spreads and cools on its own clock. A central function cannot sense it in time, and a quarterly dashboard is a map room. The design that fits the problem is the one mission command described a century ago: intent and standards held at the centre, including, yes, the master-data grammar, and ownership of the data, the intelligence and the action pushed to the frontline, with agents as the force multiplier that finally makes frontline analysis effortless rather than aspirational.
The commander of the patch
So picture the role at the end of all this. The account executive is no longer a producer of slides and a reciter of pipeline stages. They are the commander of their patch: the owner of its data, accountable for the integrity of its signal; the director of a swarm of agents that produce the material and work the cluster at scale; the reader of a contagion they can finally see; and the keeper of the only things that never decentralise, the judgement about what is true and worth spreading, and the human presence to carry it into the room. The centre stops being the factory and becomes what a good headquarters is: the holder of intent, the setter of standards, the guarantor that the whole field is playing the same game, and then it gets out of the way.
That is the shape of the job that is coming. MEDDPICC taught us what to look for. Data taught us to see it. Agents let us act on it at a scale that was unthinkable a decade ago. And ownership, pushed to the edge where the context and the stakes both live, is what finally lets the people closest to the customer command the machinery instead of queuing for it. For forty years the framework asked one honest question, is the belief real? We can answer it now. The work that is left, and it is the best work there is, is to decide which beliefs deserve to spread, and then to spread them: at the edge, at scale, by the people who own the ground.
Simon Brender spent 25 years building and leading enterprise B2B go-to-market across Asia, Europe and the Middle East, rebuilding DataRobot's go-to-market across Asia and growing Protegrity's APAC business from the first person in the region. He now runs Celerio, building data-native revenue infrastructure for founders. He learned MEDDPICC the hard way, from the people who wrote the book on it.