Blog · Published · Updated
Matching work descriptions to your estimating catalog
Short answer
An estimator receives a list of works described differently from the company's catalogue and has to find the matching items. AI suggests entries with the description, unit and scope next to the source row, and the estimator approves the match. When the codes are shared, an import is enough. A matching app costs from €3,000 net at Syntalith and pays back in 1.5 months at 2,000 rows a month at 2 minutes each.
How we workDetails ↓
- 01Call30 minutes with an engineer, free
- 02PrototypeA working system on your process and data, free
- 03ImplementationTarget and fixed price in the contract; 50% at signing
- 04Acceptance50% once the agreed target is met
- model on your server or in the cloud in the EU
- quoted after the review
- same work in application with EU cloud model
- from €3,000
- task, data and equipment analysis
- €0
We belong to the OpenAI and Anthropic partner programmes and learn through their official courses. What this means ↓
Author: Syntalith · 3 min read
Statistics
Search with AI: statistics 2026
1.5 mo.
payback of a €3,000 app at 2,000 estimate rows a month: €2,327 of searching by hand, €358 with suggestions and checking. Scenario.
49%
According to Anthropic (2024), combining contextual embeddings with BM25 search cut failed retrievals by 49%.
10.8%
According to Eurostat (2025), 10.8% of EU construction companies used AI, the lowest share of any sector.
20.0%
According to Eurostat (2025), 20.0% of EU companies with at least 10 employees used AI.
50%
According to the European Commission (2026), half of EU SMEs and mid-caps name high investment costs as a barrier to AI.
Source: European Commission, Flash Eurobarometer on AI (2026) ↗
Figures checked: 22 October 2026
The word „painting” alone is not enough
The estimate you received includes painting walls in square metres and painting railings in running metres. The company has separate entries for both jobs. The names are similar, but linking both rows to the item for walls would lose the difference in the type of work and the unit.
Source row
the description from the list
The original text and unit from the estimate received.
Catalogue suggestions
description, unit, scope
Several of the nearest entries to compare and open.
Decision
approve or clarify
A row that is too general stays for clarification, and does not get the nearest name.
Matching units helps reject a wrong suggestion, but it does not settle the match. The tool helps find recorded works, without deciding whether materials or methods are equivalent.
When an import or a search engine is enough
The same checked codes in the list and the catalogue
an import or a code search
€0
Fixed differences in naming
a list of abbreviations and synonyms, shared approved links
a small scope
Descriptions change between projects
meaning-based search with AI
from €3,000
Also check whether your current program lets you narrow results by unit and type of work. A model is worth comparing when descriptions change between projects and searching for the exact phrase gives few useful results. It helps find a similarly described scope despite a different word order or vocabulary, and the result is a list of suggestions with the original row beside it.
What justifies adapting the model
First test an available search model on the descriptions the team struggles with. If the right entries are easy to find among the suggestions, extra training is not needed. A model will not fill in a missing scope description from a short name alone. Google describes tuning embedding models to a domain on labelled examples, and you compare its usefulness for the catalogue of works with the base model. According to Anthropic (2024), combining contextual embeddings with BM25 search cut failed retrievals by 49%, so improving search alone is often enough.
What it costs and when it pays back
- Today
- 2,000 rows × 2 min = 66.7 h (€2,327)
- With the system a person checks the output in 15% of the time
- €349
- Model, 1 call per case, Claude Sonnet 5.5
- €9
- Cost of one case
- €1.16 today, €0.18 with the system
- Saving
- €1,969 a month
Payback of €3,000
1.5 months
The volume, the time and the checking share are the scenario's assumptions. €34.90 an hour is the EU average hourly labour cost for 2025 according to Eurostat. Model: 1,500 input tokens and 200 output tokens per call at Anthropic's rates, at the ECB rate of 1.1205 USD per euro on 22 October 2026.
Assessment on the next project
- Whether the right entry is among the suggestions
- How many wrong items the estimator rejects
- Which descriptions stay for clarification
- A description for which the catalogue has no item
- The whole work compared with your current search, corrections included
When the catalogue changes, keep a record of which item an approved link leads to, because a change of description or unit needs a review. Syntalith offers an app for matching work descriptions, in which the estimator sees the scope and the units, opens the source and saves the approved link. For the conversation, bring one item that regularly takes time to find.
