From bottleneck to standard process: how a system makes your quoting scalable

In most mid-sized manufacturing companies, quoting hangs on one or two experienced people. That works — as long as demand is steady and no one is out. But this is exactly where an underestimated problem lies: a process tied to individual people cannot scale.

More requests then don’t mean more revenue, but longer waits. And in peak times — seasonal spikes, or when a major customer sends several requests at once — quoting becomes the bottleneck.

The previous step was making the hidden costs and the locked-up tacit knowledge visible. This step answers the next question: how does the specialist bottleneck become a standard process the whole team can carry?

Why individual knowledge doesn’t scale

Even when historical data exists — in ERP systems, Excel files or project folders — it is rarely prepared in a way that makes it directly usable in the quoting process. The experienced employee searches archives, compares from memory, combines information from different sources. It works — but only for them.

Every new request effectively starts from zero. When several arrive at once, the rest wait. The company can’t increase its capacity without hiring new specialists and training them for years — time many companies don’t have.

“More requests then don’t mean more revenue, but longer waits.”

The way out is therefore not another specialist, but a system that makes the existing knowledge usable for everyone.

A system that works on two levels

Modern systems automatically identify similar past projects based on historical project data and propose them as a basis for calculation. The technology works on two levels — and it’s exactly this combination that matters.

Rule-based logic

The structured part is handled by fixed rules: “material + size + quantity = cost frame”. Material prices, hourly rates, surcharge factors — everything that can be clearly parametrised runs on transparent logic. In practice, this rule-based part already covers the majority of the process.

AI-supported analysis

The unstructured part is handled by AI. Intelligent pattern recognition from previous projects evaluates what can’t be pressed into fixed rules: project notes, risk assessments, the particularities of past orders. Exactly where the experienced estimator’s gut feeling used to be required, the analysis of historical data now delivers a well-founded suggestion.

The principle: “the human decides, the system prepares”

One thing matters here: this is not about full automation. The expert reviews and approves. But instead of starting from zero, they start with a well-founded proposal based on the experience of all previous projects.

The system suggests similar projects, calculates costs and identifies risks. The employee reviews, adjusts and approves. The technology takes over the bulk of the groundwork — the final decision stays with the human. Professional control is preserved, while the time-consuming routine falls away.

What changes as a result

Once the calculation knowledge lives in the system instead of in individual heads, the whole process changes its statics:

• No longer one or two employees can create quotes, but every qualified colleague on the team — supported by a system built on the company’s accumulated knowledge.
• Quote quality becomes consistent and traceable, because everyone calculates on the same basis.
• Peak times lose their threat, because capacity no longer depends on the availability of individuals.
• The company responds quickly and professionally — no order is lost because no one is available to calculate.

The bottleneck becomes a standard process. And “we can’t handle more requests” becomes “we can grow without scaling up headcount”.

What you can do

Building such a system doesn’t start with software, but with an honest inventory: which historical project data already exists — and in what form? Where does the calculation knowledge sit today? And what share of your quotes even follows a pattern that can be systematised?

Only on this basis can you decide where rule-based logic is enough and where AI delivers real value. In most cases the effort is far smaller than feared — and the first system-supported quotes often emerge within a few weeks.

Go deeper: the full whitepaper

This article is part of our series on the whitepaper “Quoting in the Mittelstand: From the 75-minute chaos to the 15-minute standard” (in German). It starts with the article The hidden costs of your quoting process. What such a transformation looks like in practice is shown in our case study from plastics thermoforming.

Download the whitepaper for free

About the author: Oliver Bührer is Managing Director of SimplifieD Solutions GmbH. With his team, he supports mid-sized companies in mechanical engineering, construction and manufacturing in adopting AI and automation solutions — with over 30 successfully delivered projects.