Excel or quoting AI? How to price bills of quantities

As of 8 November 2026. Resource type: method comparison (Excel workflow versus specialised quoting software). Not a vendor comparison. Figures come from the linked sources; Scalera statements come from the product page, FAQ and published case studies.

What does quoting a bill of quantities in Excel usually look like?

Like one workbook per tender: a row per item, a look-up in your own catalogue, and unit price plus total copied into the line. A German Leistungsverzeichnis describes the overall works as partial items, often called Positionen, and is laid out as a table (Wikipedia, Leistungsverzeichnis (German)). Internationally this is a bill of quantities, which itemises materials, parts and labour so that bidders price the same scope (Wikipedia, Bill of quantities).

Which columns does a bill of quantities have?

Typically an item number, quantity, unit of measure, long and short text, a unit price and a total price as quantity times unit price. Hierarchical group levels such as lot, trade or heading are added on top (Wikipedia, Leistungsverzeichnis (German)). A quoting workbook usually mirrors this structure one to one.

How does the price get into the row?

By matching against the article master. In the spreadsheet, staff search by description or article number, copy prices or use look-up formulas. That works while texts are unambiguous, the catalogue stays manageable and one person owns the file. As soon as tender wording diverges from catalogue names, or several variants come into play, the look-up turns into interpretation.

When is Excel the right tool?

When the task is small, one-off and well structured – not when you need to match a long or scanned bill of quantities against a very large article master.

Which tasks is Excel still good for?

• Recalculating a few items. A short recosting or a variant calculation belongs in a sheet you can adjust quickly.
• Pivot and plausibility checks. Totals by heading, material group or price band are easy to inspect in a pivot table.
• A review sheet for specialists. An export for a second pair of eyes remains a strong use – regardless of where the matching was done.

In this comparison Excel is a method, not an opponent. Keeping the spreadsheet for review and special calculations is entirely reasonable. The question is which steps it should carry, and which it should not.

Where does the spreadsheet workflow break down?

As soon as the document is unstructured, the catalogue becomes very large, or several people maintain parallel versions. Effort then grows, and mistakes stay easy to miss.

What happens with PDFs and scans?

The items first have to be transferred into the sheet. Retyping or copying piece by piece costs time and introduces transfer errors, especially in long documents. Specialised software reads digital and scanned PDFs directly, as well as Excel or CSV workbooks, GAEB files, XML and incoming messages (Scalera, FAQ). The product page also lists NPK and ÖNORM as tender standards (Scalera, Product Matching).

Why does catalogue matching get difficult?

Because a look-up formula does not show match confidence and does not explain differences between the tender text and the catalogue name. At Frauenthal the master data covers 1.6 million articles; manual processing of a bill of quantities took 4.5 hours and fell to 37 minutes with AI matching (Frauenthal case study). Match types and how to read confidence scores are covered in the product-matching guide.

Why do version risks appear?

Because logic and data live in the same file. Changes do not necessarily leave an audit trail. The European Spreadsheet Risk Interest Group notes that the majority (over 90%) of spreadsheets contain errors and that operational models are rarely tested; about 50% of models used in large businesses have material defects (EuSpRIG, Research and Best Practice).

Ray Panko summarises audits of 163 operational spreadsheets (Known 9): errors were found in 84%. The stricter core set of 55 spreadsheets (Core 5) showed important errors in 91%. In experiments, inspectors caught only about 60% of the errors present (Panko, Audits of Operational Spreadsheets).

Does a workbook learn from corrections?

No. A corrected match stays inside that particular file. The product page instead states that product data is imported once and the knowledge is retained across teams and sites (Scalera, Product Matching). The same article master is then available for every tender, instead of travelling in email attachments.

What does specialised quoting software do differently?

It handles import, matching against the article master and structured export; technical approval stays with the team. Scalera extracts items and matches products with an agent trained on construction (Scalera, Product Matching). This comparison is not a chatbot test: general-purpose language models are covered in ChatGPT for bills of quantities and quotes.

Which documents can the software read?

According to the FAQ: digital and scanned PDFs, GAEB, Excel or CSV workbooks, XML files and incoming messages with an attachment (Scalera, FAQ). The product page adds NPK and ÖNORM (Scalera, Product Matching). Excel appears here as an input format, not as a substitute for the quoting method. The DACH exchange formats are placed in context in the guide GAEB, ÖNORM and NPK.

How is matching different from a look-up formula?

The software matches on the technical specification, the intended use and the relevant standard, rather than on keywords alone (Scalera, Product Matching). Each suggestion comes with a rationale and a confidence score; missing accessories are flagged (Scalera, Product Matching). A look-up formula such as VLOOKUP either returns a hit or it does not – with no indication of how well the text and the article fit.

What do published examples show?

They show time saved on matching, not that spreadsheets are unnecessary. ACO reports that handling a 300-page specification dropped from 30–45 minutes to 10–15 minutes, with a five-week start (ACO case study). Frauenthal reports 90%+ precision on structured bills of quantities and 7.1 times faster processing (Frauenthal case study). A calculation of the economic benefit is in The hidden cost of manual tender processing – we do not rebuild that calculator here.

The question is not Excel or software. The question is which step belongs in the spreadsheet – and which does not.

Which tasks suit Excel, specialised software, or both?

The fit depends on the task, not on a blanket preference. The overview below is a recommendation, not a true/false grid against a tool.

Recommendation based on the methods described; no claim about individual vendors. As of 8 November 2026.

How do the spreadsheet and quoting software work together?

One after the other, not against each other. The software reads the tender, suggests articles and records the rationale and confidence. Your team reviews the suggestions and approves them. For colleagues who prefer to read in columns, you export a review sheet and feed the checked result back.

The structured path into ERP, CPQ or CRM is a file export or an interface – CSV, XML, GAEB and API are named; for ERP there is also the SAP IDoc format and REST (Scalera, Product Matching · Scalera, FAQ). Details of the handover are in ERP integration for tenders. Excel then remains the review and calculation sheet, while the article master, version and handover live in a shared process.

Key takeaways

• A bill of quantities is tabular: item, quantity, unit, text, unit price, total price.
• Excel is suited to recosting, pivot checks and review exports.
• Limits appear with unstructured documents, very large catalogues, versions and no match confidence.
• Audits found errors in 84% of 163 operational spreadsheets; inspectors caught only about 60% of errors (Panko).
• Specialised software imports formats, matches against the article master and hands over structured data; approval stays with the team.
• The two tools can run in sequence: matching in the software, optional review in the spreadsheet, then the ERP.

Next steps: In your last quote, mark which steps lived in the spreadsheet and which involved matching, versioning or handover. Test three typical tenders using the pilot test plan or see the flow in action: Book a demo. A broader framing is in the guide to automating proposals.

Frequently asked questions

Is Excel inherently unsuitable for bills of quantities?

No. Excel works well for compact recalculations, pivot checks and review exports. Limits appear when unstructured documents, a very large article master or several file versions are involved.

When should I move to specialised quoting software?

When you regularly match long or scanned bills of quantities against your article master and want to pass the result on without retyping. How to calculate the benefit is covered in The hidden cost of manual tender processing.

Can I use Excel and quoting AI side by side?

Yes. The software handles import and matching; your team reviews the suggestions and can use a spreadsheet export as a review sheet before data goes into the ERP.

Does Scalera process Excel files?

Yes. According to the FAQ, Excel and CSV files are among the input formats, alongside PDF (including scans), GAEB, XML and emails (Scalera, FAQ). How Excel sits next to GAEB, ÖNORM and NPK as an exchange format is explained in the guide GAEB, ÖNORM and NPK.

Does the sales office still need to review suggestions?

Yes. Each match appears with a rationale and an indication of how sure the model is (Scalera, Product Matching). Your team gives the approval.

Is this the same comparison as ChatGPT versus quoting AI?

No. The ChatGPT article contrasts a general-purpose chatbot with specialised quoting AI. This page compares a spreadsheet workflow with specialised software. More: ChatGPT for bills of quantities.

How do I test a switch in practice?

With real tenders and clear review criteria. The pilot test plan describes how to run the trial; a broader selection aid is in the guide to automating proposals.