Monday, 8 a.m.: In the inside sales inbox there is a GAEB bill of quantities with 180 line items, a photographed materials slip from the job site and an Excel list with descriptions that appear nowhere in your item master. If you want to automate quote creation, this is exactly the work you have to solve: read requests, match each line item to the right product, check specifications and pull prices from the ERP. AI software now handles this groundwork and can create quotes automatically, so inside sales only has to review and approve them.

Key takeaways

• In technical sales, the effort is not in the quote document but in the line-item work: reading, matching, checking, pricing.

• AI software reads requests in any format (email, PDF, scan, photo, Excel, GAEB), matches line items to your own catalog and creates the quote in the ERP.

• Templates, CPQ and AI solve different problems. For unstructured requests and bills of quantities, AI is the right approach.

• Inside sales remains the quality control: every quote is reviewed before it goes out, and uncertain line items are flagged instead of guessed.

• A test with your own requests shows whether it works in your product range. With Mercura, this is usually possible within two working days.

What does "automate quote creation" mean in technical sales?

In technical distribution and at manufacturers, automating quote creation means handing the recurring line-item work to software: read the request, match line items to your own item master, check technical requirements, pull prices from the ERP and create the quote in the ERP. The commercial decision stays with inside sales. Other terms for this are automatic quote creation or AI-assisted quote creation.

The term is used for very different things. For tradespeople and service providers, it usually means templates and text blocks that let you create professional quotes and send them as a finished PDF. For machine builders, it means product configurators. In technical sales, the problem lies elsewhere: the request arrives unstructured, in the language of the customer or planner, and first has to be translated into article numbers.

The following overview shows which approach solves which problem:

We describe the difference from configurators in detail in the article CPQ software or AI quote automation.

Where does the quoting process lose the most time?

Most time is lost translating the request into your own catalog: every line item has to be read, understood and matched to a product, often based on long texts, standard references and brand specifications. Writing the quote document itself is usually the smallest part of the work.

A typical inside sales workflow and what can be automated:

Steps 3 and 4 are the bottleneck. A quick calculation: for a BoQ with 200 line items, every minute per line item adds up to more than three hours of pure matching work. On top of that, the knowledge of which product fits which tender wording often sits in the heads of a few experienced employees. More on this in the article Inside sales: tasks, bottlenecks and how AI helps.

How does AI quote creation work, step by step?

AI quote creation follows a fixed sequence: the software reads the request in its original format, breaks it down into line items, compares each line item with your own catalog and the technical requirements, and creates a draft quote in the ERP. A person reviews the flagged line items and approves the quote.

Using Mercura, an AI software for quote and order automation in technical sales, as an example, it looks like this:

1. Read: Requests, tenders and BoQs are processed in any format: email, PDF, scan, photo, Excel or GAEB.

2. Match: Each line item is matched to the supplier's own catalog, including long texts. Incomplete product data is not an obstacle, because data sheets, technical documents and catalog texts are read as well.

3. Check: The requirements of the line item are checked against the product data. Compliant options are sorted by margin.

4. Source: For items that are not in stock, the software creates price requests to suppliers and reads their replies back.

5. Create: The approved quote is written to the ERP and, if needed, exported as GAEB or PDF.

6. Track: Open quotes are tracked for follow-up.

What matters is how uncertainty is handled. Line items where the match is not clear are flagged, not guessed. That way inside sales knows exactly where to look, instead of working through the whole quote a second time.

Which formats and data does the AI need?

An AI for quote creation needs two things: the incoming request in its original format and access to your own item master with product data. Perfectly maintained master data is not a prerequisite, because missing attributes can be filled in from data sheets, technical documents and catalog texts.

Relevant data and their sources at a glance:

How GAEB files are structured and which phases exist is explained in the article on the GAEB format.

Important for IT and management: the ERP remains the leading system, and prices and calculation stay there. Mercura is connected to SAP S/4HANA, SAP ECC, SAP Business One, Microsoft Dynamics 365, Salesforce, proALPHA and UniTrade, among others. Other systems are connected via an API or SFTP file exchange; for SAP, the real-time connection runs via OData, alternatively as an IDoc file. Details are on the Integrations page.

What does automated quote creation deliver in practice?

Automated quote creation gets quotes to the customer faster and makes sure more incoming requests get answered at all. Every additional request answered is a sales opportunity that would otherwise have been left sitting. How large the effect is depends on product range, request mix and data quality.

Three published customer examples from Mercura:

• Sanitär-Heinze (plumbing and HVAC distributor): up to 90% faster, from 2 days to 2 hours per quote. 100% of incoming requests are handled, compared with about 70% before. Michael Kania, Vice President Controlling: "What used to take 2 days now takes just 2 hours." (Case study)

• Muffenrohr (civil engineering supplier, 650 employees): 100% of requests handled, up to 60% faster. Managing Director Mario Hinz: "This AI will change Muffenrohr." (Case study)

• REHAU (building technology manufacturer): Product suggestions for up to 98% of BoQ line items, hit rate of up to 90%, up to 60% time savings on quotes. Martin Prigge, inside sales: "With Mercura's quoting AI, we save up to 60% of our time." (Case study)

Across all customers, the average match hit rate is 92%. It varies by customer and catalog. Mercura is used by more than 80 distributors and manufacturers in Europe and the US, rolled out across 115 locations.

When is AI quote creation not worth it?

Using AI for quote creation is not worth it if only a few, uniform requests come in or customers already order in a structured way. In that case, templates, the quote module in the ERP, a web shop or EDI are cheaper and perfectly sufficient.

Other solutions fit better in these cases:

• Low volume: If you write only a handful of quotes a month with known line items, text blocks in the ERP work well.

• Configurable products: If customers put a product together from fixed options, a CPQ configurator is the right tool.

• Structured channels: If large customers order via EDI or web shop anyway, there is nothing left to read.

• No review capacity: AI quotes are drafts. If nobody can assess flagged line items technically, clarify that role first.

• General chatbots as a substitute: Writing a quote with AI such as ChatGPT sounds obvious, but a general chatbot knows neither your item master nor your terms. It can help with individual questions, but it is not built to price a BoQ. More at Creating quotes with ChatGPT?

Even a good AI does not match every line item correctly. That is why the review step is not a concession but part of the process: the team remains the quality control before a quote goes to the customer.

Choosing quote management software: What should you look for?

When choosing quote management software and AI solutions for technical sales, look above all at three points: whether it processes your real requests in their original format, whether it matches against your own catalog including long texts, and whether it writes the result to the ERP without you having to maintain pricing logic twice.

This checklist helps in the selection process:

For Mercura: ISO/IEC 27001 certified, GDPR compliant, hosting in Europe, a separate tenant for each customer. Customer data does not train external AI models, and the connection to your systems runs only via routes your IT approves and controls. More about the product on the Quotes and tenders page.

How do you roll out the automation?

Start with a clearly defined request type and real data: typical requests from the last few weeks, the finished quotes that belong to them as a benchmark, and an export of your item master. That way you quickly see how well the matching works in your product range before you change any processes.

A proven sequence:

1. Choose a request type: for example GAEB BoQs from project business or email requests from day-to-day business.

2. Assemble test data: real requests including difficult cases, not just the clean ones.

3. Technical test: With Mercura, the test with your own data usually runs within two working days, without prior training.

4. Measure the result: match hit rate, time per quote, share of flagged line items.

5. Take the first workflow live: With Mercura, the first workflows are usually live in under four weeks.

6. Expand: more locations, request types or the capture of incoming orders, see Automate order entry.

Test it with your own data

Bring a real bill of quantities: in 30 minutes we show you how Mercura handles it in your process. Book a meeting

Conclusion

If you want to automate quote creation in technical distribution or at a manufacturer, start with the line-item work, not with the layout of the quote document. AI software can read requests in any format, match line items to your own catalog, check specifications and prepare the quote in the ERP. Inside sales reviews, decides and takes care of custom quotes for large projects instead of retyping. The fastest way to see whether it pays off for you is a test with your real requests.

Further reading

• Quotes and tenders: turn every request into a quote

• Bill of Quantities (BoQ): Structure, Item Types and Processing

• Quote-to-Order Automation: From Quote to Order Without the Gaps

• Customer story: REHAU (building technology manufacturer)

Frequently asked questions

Can AI create quotes fully automatically?

AI can largely prepare quotes in technical sales, but approval should stay with a person. At Mercura, every quote is reviewed by the sales team before it goes to the customer. Uncertain line items are flagged, not guessed.

Which software is suitable for quote creation?

That depends on your requests. For a few standardized quotes, templates or the quote module in the ERP are enough; for configurable products, CPQ software; and for many unstructured requests and BoQs, an AI quote automation that writes to the ERP.

What does quote creation in distribution involve?

Creating quotes in distribution involves reading the request, matching each line item to a product, the technical check, pricing with terms and surcharges, price requests for non-stock items, creating the quote in the ERP and following up on open quotes.

Do I need perfectly maintained master data?

No. Complete master data helps but is not a prerequisite. Mercura also reads missing information from data sheets, technical documents and catalog texts.

How accurate is the matching of BoQ line items?

At Mercura, the average hit rate is 92%, depending on customer and catalog. REHAU receives product suggestions for up to 98% of BoQ line items with a hit rate of up to 90%.

Which ERP systems does it work with?

Mercura works with SAP S/4HANA, SAP ECC, SAP Business One, Microsoft Dynamics 365, Salesforce, proALPHA and UniTrade, among others. Other systems are connected via an API or SFTP file exchange.

Is our data used to train AI models?

No. At Mercura, customer data does not train external AI models. Every customer has a separate tenant with hosting in Europe, and Mercura is ISO/IEC 27001 certified.