94% of buyers use AI: rethinking B2B sales
The quiet shift in B2B purchasing
B2B purchasing has changed fundamentally. While manufacturers and wholesalers are still debating how to digitalise their sales processes, their customers have long since put new tools to work. The study “Autonomous Commerce Shift 2026”, conducted by ONE8Y on behalf of Emporix, shows that 94 per cent of B2B buyers already use AI tools such as ChatGPT or Microsoft Copilot for product research and purchase preparation. 332 B2B suppliers and 168 heads of purchasing were surveyed in March 2026 – the findings are unambiguous.
This development puts one central question to classic B2B sales: how do manufacturers reach buyers who no longer research on the supplier's own website but put their questions straight to AI systems? This article looks at four concrete fields of action for mid-sized manufacturers and wholesalers who want to bring their sales strategy into line with the new reality.
Why AI-supported research is changing B2B sales
The figures in the study speak plainly. 94 per cent of the buyers surveyed use AI tools to find product information, compare specifications and assess suppliers. That has a direct effect on the digital visibility of B2B suppliers.
The key findings of the study:
94% of B2B buyers use AI tools for product research
74% abandon the purchasing process if it is too complicated
53% of suppliers name their IT landscape as the biggest strategic hurdle
The consequence: anyone who does not show up in the answers given by ChatGPT, Copilot or other AI systems is simply not found. Classic SEO strategies are no longer enough once buyers stop starting their research on Google and go straight to a chatbot.
For mid-sized businesses this is a double challenge. On one side, the technical infrastructure to supply product data in a form AI can read is often missing. On the other, many sales processes are still built around personal contact and manual handling – a model that is increasingly reaching its limits.
First answer: structure product data so that AI can read it
AI systems work with structured data. Marketing copy full of adjectives and vague claims does not get you anywhere. What counts are clear facts: technical specifications, dimensions, materials, standards, certifications.
Concrete measures for structured product data:
Standardise technical data sheets: create uniform formats for all product groups
Name attributes clearly: instead of “high-quality workmanship”, concrete details such as “Material: stainless steel 1.4301, DIN EN 10088”
Use structured data: implement Schema.org markup for products
Introduce a PIM system: Product Information Management as the central data basis
The challenge often lies in data quality. Many mid-sized manufacturers keep their product information in Excel spreadsheets, ERP systems or even printed catalogues. That data is readable for people, but barely usable for AI systems.
A Product Information Management system (PIM) is the remedy here. It centralises all product data, allows uniform structures and makes sure that technical information is consistent and complete. Importantly, investing in data quality pays off several times over – not only for AI visibility, but also for internal processes and classic sales channels.
Second answer: become present in AI answers
In 2026, digital visibility means more than a well-ranking website. It means appearing in the answers AI systems give when potential customers look for products, solutions or suppliers.
Strategies for AI visibility:
Make prices and availability publicly accessible: AI systems favour concrete information
Publish technical documentation online: data sheets, CAD models, certificates as downloads
Build structured FAQ sections: give typical buyer questions clear answers
Publish customer reviews and references: first-party data as a trust signal
Many B2B suppliers hesitate to show prices online. The worry is that competitors could use the information or that customers could order without negotiating. The reality is different: 74 per cent of buyers abandon the process when it is too complicated. Anyone who offers no price indication loses those prospects to competitors who operate more transparently.
A pragmatic solution: show list prices or price ranges publicly and hold customer-specific terms in the protected customer portal. Room for negotiation is preserved, while the first hurdle for buyers comes down.
Third answer: self-service for well-informed buyers
When 94 per cent of buyers start their research with AI tools, they enter the purchasing process already very well informed. They know the specifications, have compared alternatives and know what they need. What they do not want: long waits for quotations or queries about standard information.
What modern B2B e-commerce solutions have to deliver:
Real-time stock display: availability visible directly in the shop
Customer-specific prices: after logging in, every customer sees their own terms
Fast reordering: order history available, repeat orders with a single click
Configurable products: variant selection without a query to sales
The classic objection runs: “Our products are too complex for self-service.” The study suggests otherwise: complexity is not an argument against digital processes but a reason to design them particularly well. A well-structured customer portal with filters, technical comparison tables and clear product data reduces the complexity for the buyer.
Self-service does not mean replacing personal contact. It means automating routine processes and freeing up sales resources for the cases that genuinely need advice. A buyer ordering the same screws for the third time does not need a field sales visit. A customer who wants to introduce a new product line does.
Fourth answer: redefine sales roles
The shift in purchasing behaviour calls for a realignment of sales. When buyers research on their own and place standard orders themselves, the role of field sales and internal sales moves.
New priorities for B2B sales:
Advice instead of order taking: focus on complex enquiries, technical special cases, new business
Data-driven customer analysis: use customer insights to spot needs proactively
Omnichannel sales: seamless transitions between digital self-service and personal advice
Sales transformation: enable teams to use digital tools and interpret customer data
The biggest hurdle often lies not in the technology but in the organisation. 53 per cent of the suppliers surveyed name their IT landscape as a strategic barrier. Behind that, though, there are usually organisational questions: who is responsible for product data? How are prices maintained? Which processes run digitally, which stay analogue?
A successful sales transformation starts with clarity about roles and responsibilities. Field sales does not have to turn into an IT expert, but it should understand which data the system records and how that data can be used for customer analysis. Internal sales needs access to real-time information on stock, delivery times and customer history in order to respond quickly and competently.
The role of first-party data in data-driven sales
Anyone who builds self-service channels gains valuable first-party data. Every search in the shop, every product view, every order gives an insight into customer needs. The quality of this data far exceeds external market research, because it shows real behaviour rather than stated intentions.
Using first-party data in B2B sales:
Personalisation: product recommendations based on order history and sector
Optimise pricing models: demand-based adjustments, volume prices, campaigns for specific customer segments
Sales transformation: field sales receives insights into customer activity and can advise in a targeted way
Customer analysis: spot churn risks early, identify cross-selling potential
The challenge: many mid-sized companies collect data but never evaluate it systematically. A simple way in: monthly reports on the most-searched products, the most frequent drop-offs in the ordering process and the most active customer segments. This information helps to improve range, prices and communication continuously.
Overcoming technical hurdles pragmatically
The study shows that 53 per cent of suppliers see their IT landscape as the biggest hurdle. Long-established ERP systems, processes that have grown up over the years and missing interfaces make rapid change difficult. Even so, there are pragmatic routes.
Building up step by step instead of a big-bang project:
MVP approach: start with a minimum viable product that makes the first sales possible
Modular integration: connect systems step by step instead of switching everything at once
Use external expertise: bring in partners with experience of ERP, PIM and shop integrations
Prioritise quick wins: implement the measures that show an effect quickly
A common mistake: companies want the perfect solution from the outset. That leads to projects lasting months, high investment and often to frustration when requirements change during implementation. Better: start with a working core, gather initial experience and then extend step by step.
What this means for manufacturers and wholesalers
The change in B2B purchasing is not a distant future but current reality. Anyone still waiting for customers to call or write of their own accord is losing market share to competitors who are better set up digitally.
The good news: getting started does not have to begin with investment running into millions. More important than the perfect system is the willingness to begin. Structure product data, make initial information available online, build a simple self-service channel and develop sales along with it step by step.
The four fields of action – structured product data, AI visibility, self-service and sales transformation – interlock. Clean data can be made available to AI systems. Offering self-service produces valuable customer insights. Realigning sales creates capacity for strategic account management.
The shift calls for no revolution, but for a clear strategy and consistent execution. Companies that act now secure an advantage in a market that is changing fast.
