From FR Advisory.
The author spent twenty years running commercial organisations inside software and private equity portfolios. He no longer works in the technology industry, holds no position in any company named in this series, has no clients in it, and is selling nothing to anyone in it.
Start in the room, because that is where this gets decided.
A seller is in a renewal conversation. They are asking for about seven and a half per cent more than last year. The product has artificial intelligence in it now, which is the reason given for the increase. The buyer asks a reasonable question, which is what the extra money buys.
And the honest answer is a tier. Not more work done. Not a measured result. A tier, with features in it, priced higher than the tier below it.
That is the entire problem with this year, and everything underneath it, pricing, metering, compensation, revenue architecture, exists to explain why that seller is losing a conversation they should be winning.
What the buyers actually asked for
Buyers have been unusually clear about this, and it is not ambiguous.
In the first half of this year, Futurum Research asked enterprise software decision makers how they want to pay. 43% said consumption. 27% said outcomes. Fewer than one in five still prefer paying per person.
Seven in ten want the bill to move with something real. Either the work the software does, or the result it produces.
That is a buyer telling you, out loud, that they will pay more if the number moves with something they can see. It is the easiest brief a commercial team has been handed in twenty years.
What the industry shipped instead
Now the other side of the same market.
Teneo surveyed three hundred United States software vendors between late December and late January. They asked how AI is being paid for.
Usage based, 38%. Hybrid, 33%. User based, 26%.
Outcome based: 2%.
Then they asked how the AI is packaged. 42% put it in a new, higher tier. 27% embedded it in the core product. Together, 69% of vendors answered the arrival of agentic AI by moving a feature list up a price page.
Seventy per cent of buyers asked for a meter. Two per cent of vendors sell one. Sixty nine per cent shipped a tier.
That gap is not a marketing problem or a product problem. It is the reason your renewals are harder this year than last.
The number that proves it is already costing you
Teneo asked vendors which stage of their sales process got worse in 2025.
The answer was pricing and negotiation, cited by 31%. More than discovery. More than closing. More than anything else in the cycle.
Of course it was. You sent people into negotiations with a bigger number and no new argument.
The same survey has vendors planning an average 7.6% increase on like for like renewals this year, with 51% expecting more customer pushback than usual. So the people who just told you the pricing conversation is the part that broke are being asked to do more of it, harder, with the same materials.
And then the number that should stop a board meeting. Around 70% of those vendors reported higher turnover in their go to market teams last year, and four in five said that turnover limited their ability to hit new business targets.
Sellers do not leave because the product is hard to explain. They leave when the pitch stops surviving the room, because that is when the money stops.
The invention was never the machine
Here is where it is worth being a student of history, because this exact mistake has a name and a date.
Thomas Edison built the better system. Pearl Street Station opened in 1882 and it worked. He had the generators, the lamps, the distribution, the patents. By any technical measure he had won.
Electricity still did not become an industry for another twenty years, and the man who made it one was not an inventor at all.
Samuel Insull was Edison's secretary. He left for Chicago Edison in 1892 and inherited a business selling light to rich people and factories, at a price that made it a luxury. The technology was not the constraint. The way it was sold was.
So Insull went to England, and he went to a seaside town.
Arthur Wright was the electrical engineer at Brighton. He was not famous, he was not an inventor of the technology, and he was running a municipal lighting operation on the south coast. What he had worked out was that every customer cost his company in two separate ways. There was the cost of standing ready to supply whatever they might draw at their peak, which existed whether they used it or not. And there was the cost of the electricity they actually consumed.
So Wright charged for both, separately. One rate covering the capacity held available for you. A much lower rate for everything you then used. At Brighton those two rates were threepence and a penny.
Insull interviewed him, learned the method, and took it back to Chicago.
That is the whole story. The dynamo was Edison's. The tariff was Wright's. And it was the tariff that turned electricity from a luxury good into the largest industry of the twentieth century. By 1930, seventeen of Insull's companies held about twelve per cent of all utility investment in the United States.
Nobody remembers the man who worked out how to charge for it. Everybody remembers the man who built the machine. The money went to the first one.
Why that specific invention worked
Look at what Wright's two charges actually solved, because it is the same problem sitting on your price page right now.
A single flat charge punishes the light user and undercharges the heavy one. A single usage charge leaves the supplier carrying the cost of capacity that someone might call on at any moment and never pay for. Neither one matches how the thing is really consumed.
Two charges match it. The customer with a predictable, modest draw pays little. The customer who wants the capacity there at all times pays for the privilege of it being there. The supplier's costs and the customer's bill finally point in the same direction, and the customer can forecast it.
Now say that in the language of this year. One charge for the capacity you hold open. One charge for the work actually done. Predictable for the buyer, variable where the cost is variable, and both halves visible.
That is a hybrid model. It is a hundred and thirty years old.
And it is currently ahead of most of this market. Kyle Poyar's survey of two hundred and thirty software and AI companies in April and May this year put hybrid adoption at 37%, up from 25% a year earlier. Moving fast, and still a minority.
The industry is slowly rediscovering, by trial and error, a structure a Brighton engineer shipped in the 1890s.
Why sensible people keep shipping the tier
It would be easy to call this laziness. It is not, and that misreading is why most pricing advice goes nowhere.
A tier is a price change. It needs a web page, a packaging decision and a sign off. Nothing else in the business has to move.
A meter is not a price change. A meter is a product change, a billing change, a contract change, a revenue recognition change and a compensation change, all at once, and all of them have to land in the same quarter or the model does not work.
You need to count the thing you are charging for, accurately enough to put on an invoice a customer will dispute. You need a billing system that can take that count. You need a finance team willing to forecast a revenue line that moves. You need a seller who is paid for something other than the size of the signature. And you need a renewal owner whose number depends on the customer actually using it.
A tier requires one meeting. A meter requires the operating model.
So the tier wins, every quarter, in every company, for reasons that are individually defensible and collectively fatal. That is a system working exactly as designed, which is why naming it as stupidity gets you nowhere and naming it as structure gets you somewhere.
What the 2% know
Two per cent of vendors charge for outcomes. They are worth looking at, because they are not braver than everyone else. They are structurally different.
What they have in common is a countable unit that the customer already recognises as the job. A resolved support conversation. A completed case. A produced document. Not a token, not a credit, not a compute minute. A thing the customer was already paying a person to do.
That is the test, and it is the same test Wright passed. Can you count the unit, can the customer recognise it without being taught, and would they have paid for that unit anyway.
If the answer is no, outcome pricing is not available to you yet, and chasing it will produce an attribution argument you cannot win in front of an auditor. If the answer is yes, and you are still selling a tier, you are leaving the Insull position open for a competitor.
Three questions for your next revenue review
What does one unit of your AI cost, in a unit the customer recognises. If the answer involves explaining what a credit is, you do not have a meter. You have a currency, and the customer cannot forecast their bill.
What is the seller paid when consumption goes up after signature. If the answer is nothing, your sales organisation will keep selling the biggest tier it can, which is the behaviour you designed.
If a competitor published a price per completed job tomorrow, what would you say. You have until someone does. The 2% are not a rounding error. They are a bridgehead.
The position
Seven in ten buyers asked to pay for consumption or results. Two per cent of vendors sell that. Sixty nine per cent answered with a higher tier, and then asked their people to go and defend a seven and a half per cent increase with it.
Pricing and negotiation was the stage that broke last year. Seventy per cent of those companies lost more sellers than the year before. Neither of those is a coincidence, and neither is a sales execution issue.
Edison built the better machine and nearly lost the industry. Insull copied a tariff from a municipal engineer in Brighton and took it.
Your competitors are all building the machine. Nobody is building the tariff.
The Agentic Commercial Model. agenticcommercialmodel.com
Sources
- Futurum Research, 1H 2026 Enterprise Software Decision Makers survey, reported 12 May 2026: 43% of buyers prefer consumption-based models, 27% favour outcome-based structures, fewer than one in five still prefer per-user pricing. Futurum does not publish the sample size or methodology for this survey and the full report is behind a subscription, so treat these as directional rather than precise
- Teneo, 2026 B2B Software Vendor Survey (fourth edition), 300 United States software vendors, fieldwork 23 December 2025 to 22 January 2026: primary AI value capture is usage-based 38%, hybrid 33%, user-based 26%, outcome-based 2%; AI packaging is a new higher tier 42%, embedded in the core product 27%, add-on 17%, standalone 14%; 42% introduced AI-specific pricing in the past year, mostly incrementally; among vendors concerned about seat reductions, 14% plan to move off seats completely, 40% plan hybrid models alongside seats, 28% are undecided and 18% have no plans to change; expected like-for-like renewal increase averages 7.6%; 51% expect more customer pushback than in 2025; pricing and negotiation was the sales stage that deteriorated most in 2025, cited by 31%; around 70% reported higher go-to-market turnover year on year, with four in five saying it limited new-logo attainment. The report does not state its sampling method, weighting or margin of error
- Kyle Poyar, The 2026 State of B2B SaaS and AI Monetization Report, poll of 230 B2B software and AI companies, April to May 2026, published 13 May 2026: hybrid pricing adoption at 37%, up from 25% twelve months earlier; three in four software companies changed pricing or packaging in the past year; 29% currently use AI credits with a further 33% planning to add them within 6 to 12 months. Author-run poll, no stated sampling method
- Arthur Wright, chief engineer at Brighton, devised the two-part tariff charging separately for connection capacity and for electricity consumed, at rates of threepence and a penny per unit. Sources: Grace's Guide to British Industrial History, entry for Arthur Wright (c.1858–1931); Electrical History Society supplement on Brighton, which records that Samuel Insull visited Brighton, interviewed Wright, learned the tariff method and took it back to Chicago. The two sources disagree on when the visit took place, the Brighton supplement placing it on a UK holiday in the 1880s against Insull's move to Chicago Edison in 1892, so no date is claimed here
- Institute for Energy Research, Energy Pioneers: Samuel Insull: Insull joined Chicago Edison in 1892; created a two-part rate combining a demand charge for peak capacity with a volumetric charge for power used; raised load factor by adding off-peak loads such as streetcars; by 1930 seventeen of his companies held total capitalisation of $2.2 billion, approximately 12% of national utility investment
- Edison's Pearl Street Station began operating in 1882 with six 100-kilowatt generators (Institute for Energy Research)