A wave of creation: what AI produces in your team and at your customers

A customer used to email a short question. Now they send an eighty-point questionnaire, prepared and drafted with AI. It cost the customer very little effort. Your inside sales team, though, has to answer every point and check every answer before the quote goes out.

You have probably noticed it yourself: a quote that used to take you half a day now comes together much faster with AI, neatly formatted and fluently written. What holds for your quote also holds for your colleagues' quote variants, your customers' questionnaires and your competitors' price lists. A swelling wave of AI-generated information is on its way. Even if you would rather wait before working with AI, you will not be able to escape it.

In the previous part you saw how a commercial organisation fits together: three levels, from the commercial policy down to the daily work of every employee, and six steps from the first request to expansion with an existing customer. What does AI do at the bottom level, where your team and your customers produce quotes, emails and questions every day?

AI produces more and more

The pace at which AI creates more and more information has been measured well in a few places. Google, for instance, tracks how much text its AI models process.

An AI model splits text into 'tokens', fragments of a few letters. A two-page quote easily runs to a few thousand. The number of tokens therefore shows how much text AI reads and writes. In May 2024 Google processed around 9.7 trillion tokens a month. Two years later that had grown to 3,200 trillion, roughly 330 times as many, as Sundar Pichai said at Google I/O 2026.

On the internet, the balance between people and machines has meanwhile tipped. A large share of that traffic passes through Cloudflare. In June 2026, for the first time, more than half of all website requests there came from bots: 57.5%, according to Cloudflare Radar.

The growth is not levelling off either. Five tech companies now invest more in data centres than the whole world invests in oil and gas extraction, the International Energy Agency calculated. The models themselves are also becoming more capable and more autonomous. In 2023 an AI model could complete, without intervention, tasks that take a person a few minutes. In 2026 it is already handling tasks of several hours. METR's measurements show no sign of a slowdown yet. Everything points to AI creating more and more information.

More variants in your own team

That wave of information washes up at your door too. There are two reasons. The first lies in your own team.

A colleague used to make one version of a quote. To improve his chances of an order, he can now put three variants in front of his customer with little effort, for example with a different specification or a shorter lead time. Suddenly your colleague or his sales manager has to check three quotes instead of one.

That is, if your colleague still has his quotes checked by the sales manager at all. Working with AI does not always happen in the open. In a British survey by Deloitte, 31% of people who use AI at work admitted they do so without their employer knowing.

More questions from your customers

The second reason lies with your customers. The eighty-point questionnaire from the opening is no exception. Requests like these weigh on your commercial team. The more your customers work with AI, the more of them arrive.

That is especially true if your customers are large companies, because it is mainly large companies that use AI today. According to Eurostat, around 29% of Belgian companies with ten to 49 employees used AI in 2025, against 76% of companies with 250 employees or more.

Customers also use AI when they buy. For consumers this has already been measured: over the 2025 holiday season, US online shops received almost eight times as much traffic from AI tools as a year earlier, Adobe reported.

Business buyers are also turning to AI more often, mainly to research suppliers and products. In 2025 the research firm Gartner asked 645 B2B buyers how they handled their most recent purchase: almost half (45%) used AI. The same firm expects AI agents to handle more than 15 trillion dollars of B2B purchases by 2028.

That affects your digital marketer too. An AI model forms its picture of your product from the information about your company online. Your website and your product information therefore have to be readable for AI models and agents. You can read more about that in the part on how every commercial role changes.

Thanks to all that research, buyers find more potential suppliers. Their request for quotation therefore lands with more companies than before. A British injection moulder of precision parts noticed exactly that, as Kenny and Pogrebna described in Harvard Business Review in 2026. Once its customers started using AI, the company received far more requests for quotation. At first sight they looked promising. Yet more and more often they were sloppy: the application was missing, or a competitor's name was still in them. A request like that had little chance of turning into an order. An engineer still had to check the drawing and assess the material, the mould and the feasibility. More requests brought in no extra orders. Each order did take more work.

You no longer set the volume yourself

On both the buying and the selling side, creating has become cheap. Your customers' requests flow into your inside sales team. Your colleagues' quotes and variants end up with the sales manager for his signature. You no longer set that volume yourself: it grows with the AI use of your customers and of your own team.

There is no escaping that wave. But everything that gets made also has to be checked by someone before it reaches the customer. How do you keep a grip on everything that is being made? That is the subject of the next part: why checking takes more time than making, and where the work piles up.