Every time I talk to partners and consultants about AI, three worries come up. Partner managers say if we deliver faster, we earn less. Consultants say if AI does the work, they lose their jobs. Partners say if projects finish quicker, they run out of projects. These are fair worries. I want to answer them properly, and the clearest answer comes from a moment in history that looks almost exactly like ours: the arrival of electricity.
Look at the world economy over two thousand years. For roughly 1,800 of them, it barely grew. A worker in the year 1000 produced about as much as one under the Romans. Then, in the century after electricity and industry arrived, output exploded. The line is almost flat for nearly all of human history, and then it goes straight up.
This matters more than total GDP, because total output can rise just because there are more people. Output per person cannot. When it jumps, it means each person is genuinely producing more. That is productivity, and productivity is the only thing that ever made anyone richer.
Here is the detail that changes the whole story, and it is the answer to all three worries.
When electric motors first reached factories in the 1880s, productivity did not rise. For nearly three decades, almost nothing happened. Economists still call it a paradox: the motors were everywhere except in the productivity numbers.
Why? Because the first factory owners just swapped the steam engine for a big electric motor and kept everything else the same: the central shaft, the belts and pulleys, the whole layout built around one power source. They electrified the factory without changing the factory. And it did almost nothing.
The gains came only when a new generation asked a different question. Not "how do I plug electricity into my old factory," but "what would a factory look like if it were built electric from the start." The answer was a small motor on every machine, the central shaft gone, the floor rearranged around the flow of the work. That redesign, not the motor, is what finally doubled productivity. By the 1920s, electrification drove half of all productivity growth in manufacturing.
The winners were not the ones who bought the technology. They were the ones who reorganized around it.
You can see the payoff at the country level. Britain industrialized first and climbed steadily. Japan stayed flat for a century, then adopted the whole toolkit in a concentrated burst after 1950 and caught up in fifty years what took Britain a hundred and fifty.
Worry 1: Faster delivery, less money
The pie grows, it does not shrink
When electricity made factories faster, factories did not make less. They made far more, at lower cost, so more people could afford the output, so demand grew, so there was more work than before. Cheaper and faster did not shrink the market. It opened it up.
ERP is the same. Most mid market manufacturers never start a rollout because it is too slow, too risky, and too expensive. Cut the time by more than half and a whole set of projects that were never viable suddenly are. You do not split a fixed pie into smaller slices. You bake more pies. Faster delivery per project plus many more projects beats slow delivery on a handful.
Worry 2: AI takes my job
Like the farmer who learned machinery
When machinery reached farms, the farmers who learned to use it did not lose their jobs. They became far more productive and more valuable. The ones left behind were not replaced by machines. They were replaced by other farmers who used machines.
AI does the repetitive parts of a rollout: writing test scripts, drafting requirements, building training material. That frees the consultant for the work only a person can do, the client relationship, the judgment calls, the tricky configuration decisions. A consultant with AI handles more projects and harder problems than one without. The job does not disappear. It levels up. The real risk is not AI replacing consultants. It is consultants who use AI replacing those who do not.
Worry 3: Fewer projects to go around
Speed unlocks demand that was always there
This is the same point as the first, seen from the partner side. Faster does not mean fewer. It means the projects that were too slow and too costly to attempt become worth doing. The market for ERP is not fixed at today's size. It is capped by today's speed and cost. Lower both and the market grows. More manufacturers modernize, more roll out, and there is more work, not less.
How big is the market this actually opens up
Worldwide there are roughly 5 million factories. 54% of them still run on paper and spreadsheets, not even a proper manufacturing system, let alone a modern ERP. Only 8% use a commercial system built for the job. In the US alone, of the 239,265 manufacturing firms tracked by the National Association of Manufacturers, all but 4,177 have fewer than 500 employees. Almost all of American manufacturing is exactly the mid market segment that today's ERP timelines and costs have priced out.
That price and timeline problem is current, not historical. Panorama Consulting Group's 2025 ERP Report puts the average implementation at about 9 months and $450,000, with more than a quarter of projects running over budget, and manufacturing worse than most industries for both. Cut that in half, and a large share of that 54% stops being "too risky to attempt" and starts being a project that actually happens.
Sources: IoT Analytics, MES Market Report 2025 to 2031, December 2025. National Association of Manufacturers, Facts About Manufacturing, 2022 data. Panorama Consulting Group, The 2025 ERP Report.
Here is the lesson from the factory floor, applied to us. Do not bolt AI onto the old way of running a rollout. That is the steam engine mistake, and it delivers almost nothing. Redesign the rollout around what AI makes possible.
That is exactly what we are building at Infomind: a platform where the whole ERP rollout lives in one place, and agents handle requirements, testing, training and more, with the consultant steering. We are not electrifying the old factory. We are building the one that was born electric.
The partners and consultants who reorganize around this will be the Japan of our story: late to nothing, fast to everything, and far ahead of the ones who wait. The tools are here. The only question is who redesigns around them first.
People often ask us directly who loses out here. It is a fair question, and we answer it in detail in Will AI Replace ERP Consultants? The short version: it is the task that goes, not the person who does more than the task. Script-following testers, tier-1 support routing, and manual documentation writing get automated. Judgement, client relationships, and knowing what a change will break downstream become more valuable, not less.
Sources: world and country GDP per person figures (1990 international dollars) are from the Maddison Project Database, University of Groningen, distributed via Our World in Data. The Britain vs Japan comparison is our own reading of that same dataset, not a separate published statistic. The electrification productivity story, including the three decade lag and its roughly 50% share of 1920s manufacturing productivity growth, is drawn from Paul David, "The Dynamo and the Computer: An Historical Perspective on the Modern Productivity Paradox," American Economic Review, 1990. Pre 1900 economic figures are careful historical estimates, not measured accounts. ERP market figures are from IoT Analytics, MES Market Report 2025 to 2031 (December 2025); the National Association of Manufacturers, Facts About Manufacturing (2022 data); and Panorama Consulting Group, The 2025 ERP Report.