It is interesting to put together two things that happened in June.
One is a report released by Germany's Bitkom association, with a very direct title: AI agents will fundamentally change the sales model of IT services, as traditional licensing and hourly billing are becoming obsolete. The B2B market will shift toward "outcome-based pricing"—customers will only pay for the final result, such as resolving a ticket or fixing a security vulnerability, rather than paying for "how many hours this person worked."
Another piece is that Odoo released a set of data at its partner day in the DACH region (Germany, Austria, Switzerland): from January to May 2026, revenue in the German market grew 133% year-on-year, partners increased by 70%, and the number of new monthly customers doubled, officially surpassing France to become the second-largest market in EMEA. During the same period, Odoo launched a new affiliate program—earning a 10% commission on the first 12 months of subscription revenue from referred customers, with no cap.
These two things, when viewed together, are about the same thing: AI in the ERP track is transforming from a "technical variable" into a "business variable." It is not only changing how ERP software itself is used, but also how ERP is sold, deployed, and priced.
Why Bitkom says "hourly billing" is dying
Bitkom is Germany's digital industry association, and its voice carries weight. Its core assessment is not complicated: AI agents can process a large number of IT tasks in parallel, and at a much faster speed than humans. Work that used to take an ERP implementation consultant 40 hours can now be completed by a properly configured AI agent in 4 hours or even less—and the quality is not necessarily inferior.
This creates a fundamental contradiction under the hourly billing model: the more efficient the service provider, the less money they make. If I use AI to reduce implementation time by 90%, my income under hourly billing would also shrink by 90%. This conflict is unsustainable. The market will inevitably shift to "pay-for-results"—I charge for delivering a usable system, a configured module, or a streamlined business process to the client. Whether I use humans or AI is a matter of my own efficiency.
Result-oriented
Pricing Bitkom predicts: AI agents will drive B2B IT services from hourly billing to outcome-based payment
To be honest, this trend has already begun to emerge in the high-end consulting field. In 2023, Accenture was already promoting the "outcomes-as-a-service" model. But Bitkom has applied it to the more specific scenario of ERP implementation and given a time window: not five years from now, but now.
The reason is also straightforward—in high-frequency ERP implementation scenarios, the substitutability of AI agents has become clear. Demand forecasting, invoice matching, inventory optimization, anomaly detection—these are not tasks requiring human creativity, but rather pattern recognition plus rule execution. AI's performance on such tasks is already good enough, and it is accelerating toward even better.
Odoo's report card in Germany is no coincidence
Odoo's revenue in the DACH market grew 133% from January to May, while its partner base expanded by 70%—these two figures are quite eye-catching in the 2026 ERP market. My assessment is that this is directly related to the "pay-per-result" trend.
Odoo's product logic is naturally suited for this pricing shift. An open-source, modular ERP that doesn't require expensive license fees inherently has a much shorter implementation cycle than SAP or Oracle. Coupled with the upcoming Claude AI native integration and offline support in Odoo 20, implementers can deliver a complete business system in less time with fewer people. What customers see is "the project is live, inventory is accurate, and the procurement process is running smoothly," rather than "how many days this consultant has been here."
The affiliate program launched by Odoo at the same time is also quite interesting. Recommending customers earns 10% of the 12-month subscription revenue, with no cap—this is essentially a form of "pay-per-result": the referrer receives the ongoing value generated by the subscription, rather than a one-time commission. This design encourages people to keep promoting, rather than stopping after making a referral.
Microsoft Dynamics 365 is also doing something similar
At the end of June, Microsoft launched fully autonomous AI agents in Dynamics 365—these agents can autonomously record customer calls, update sales pipelines, and handle scheduling. The accompanying Agent 365 management framework allows IT administrators to centrally monitor what data the AI agents access and shut down any agents that overstep their authority at any time.
Microsoft's move indicates one thing: Major ERP vendors are also promoting the deployment model of "AI executors". It's not about implementing a system first and then gradually adding AI capabilities, but rather deploying AI agents directly into business processes from the moment the system goes live. This also means that ERP delivery and pricing models must be adjusted accordingly—you can't charge for tasks automatically completed by AI agents using traditional per-person-per-day billing.
What this means for CIOs and ERP implementers
I have organized three levels of impact, each directly related to the decisions in the second half of 2026.
First, the pricing model for ERP procurement is being restructured. In the past, selection was based on feature lists and per-user monthly unit prices. In the future, the comparison may focus on "how much labor this system can save me after implementation" and "whether the implementer charges based on guaranteed outcomes or per headcount." The risk structure of the latter is completely different.
Second, the capability model of the implementer must change. If pay-per-result becomes mainstream, implementers can no longer profit by extending project timelines, but must rely on rapid delivery and continuous optimization. This means that mastery of AI toolchains, judgment of data quality, and understanding of business processes will replace "how many people can be stationed on-site" as core competitiveness.
Third, the Odoo ecosystem has taken the lead in this trend. Modular architecture, low licensing costs, native AI integration, and rapid expansion of the partner network—these elements align perfectly within the "pay-per-result" framework. The 133% growth rate in Germany is no coincidence; it shows that the market is already moving in this direction.
A word of caution: The pay-per-result model also poses risks for the implementer—if the AI model is inaccurate, the data foundation is poor, or the business process understanding is insufficient, the cost of delivery failure is entirely borne by the implementer. This model demands genuine expertise, not a brute-force approach.
A few final words
I don't think traditional hourly billing will completely disappear by 2026, but the direction is already clear. Over the past few months, SAP has been promoting autonomous enterprises, Salesforce has acquired AI pricing company m3ter, and Odoo has doubled its growth in Germany—each with different strategies and paths, but all pointing to one thing: The business logic of ERP is shifting from "selling software + selling man-hours" to "selling outcomes."
For CIOs, the next question is not "whether to adopt AI," but "whether this ERP vendor is paid based on results or on man-hours." This distinction determines who is sharing risks with you and who is shifting risks onto you.
