Blog · Published · Updated
What did customers say in product research?
Short answer
An AI knowledge base on the product research archive finds a customer's words together with the context of the conversation, so the team can tell a request for a feature from a problem finding one that exists. At Syntalith an AI app on company documents costs from €3,000 net. At 120 research questions a month with 20 minutes of searching each, it pays back in 2.5 months.
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
- net per application, fixed price before contract
- from €3,000
- from signing the contract to acceptance
- 4–10 weeks
- free prototype on your documents
- €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
AI knowledge bases: statistics 2026
2.5 mo.
payback of a €3,000 app at 120 research questions a month: €1,396 of searching today, €212 with the app. Scenario.
49%
According to Anthropic (2024), combining contextual embeddings with BM25 search cut failed retrievals by 49%, and by 67% with reranking.
20.0%
According to Eurostat (2025), 20.0% of EU companies with at least 10 employees used AI.
Source: Eurostat, 20% of EU enterprises use AI technologies (2025) ↗
90%
According to the European Commission (2026), 90% of companies using AI say that people take the decisions, and 2% of decisions are fully automated.
Source: European Commission, Flash Eurobarometer 578 (2026) ↗
11 of 13
According to the NoLiMa study (2025), at 32,000 tokens of context 11 of 13 models fell below half of their short-context performance.
Figures checked: 22 October 2026
Did the customer need a feature or fail to find it?
During planning, someone suggests adding a data export. Someone else remembers an interview where a customer said: „I couldn't download the report”. Later in that conversation it turns out the participant did not find the existing button. In another interview a customer needed a report with a field the product really does not have. These are two different product decisions, and the same quotation leads to both.
Passages
how many were found
The number of matching comments in the archive.
Participants
who was asked
People invited to specific studies, with different questions.
Scale
a separate measure
Judging how often a problem occurs needs quantitative data or a new study.
A few matching passages do not say how often the problem occurs among all customers. The researcher reads the context of every passage used, and the product decision rests on the study it came from, with its questions and its choice of participants.
What a product research knowledge base costs and when it pays back
- Today
- 120 research questions × 20 min = 40 h (€1,396)
- With the system, a person checks the output in 15% of the time
- €209
- Model, 1 call per case, Claude Sonnet 5.5
- €2.52
- Cost of one case
- €11.63 today, €1.76 with the system
- Saving
- €1,184 a month
Payback of €3,000
2.5 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: 8,000 input tokens and 700 output tokens per call at Anthropic's rates, converted at the ECB rate of 1.1205 USD per euro on 22 October 2026.
The model is a small item here. The result depends on the time that searching and asking around take today, and on how much of it is left for checking the answer.
Before we quote a build, we run a test on 20 to 100 questions the team has already asked: we compare the app's answer with what an employee found by hand and measure the time from the question to the source. The test takes 5 working days and costs €0.
Check the tool that already holds the research
Dovetail describes searching research material and answer summaries based on the results, depending on the plan. If the team keeps its studies there, test these features on your own questions. A separate app makes sense when transcripts, notes and recordings sit in several places, or when questions join research with product usage data.
Studies in one research repository
search in the current tool
vendor licence
Transcripts and notes in different folders
an AI app on the research archive
from €3,000
A question about the scale of a problem
a quantitative study or usage data
a separate measure
Start with a product question
The first scope is the archive from the last year and the questions where the team tries to go back to earlier studies. The researcher judges whether the answer keeps the context of the words, and the product manager whether it helps the decision. Material with participants' personal data stays within the access agreed with the person responsible for research. The scope is described on the private ChatGPT for your company page.
For the first call, one question that regularly comes back to colleagues and a note of where the team looks for the answer are enough. You do not need to send confidential documents, and you get a written plan within 2 working days.
Worked example · Scenario
How we count the payback: answers from company documents
We run the same count for your documents in the free analysis.
€3,723
a month for finding answers to 800 questions in documents
Payback: 1 mo.
Our assumptions are 8 minutes to find an answer in documents or ask a colleague, and checking in 15% of the old time. €34.90 an hour is the EU average hourly labour cost for 2025 according to Eurostat, for companies with 10 or more employees. Model cost: Claude Sonnet 5.5, about 6,000 + 500 tokens per answer from documents, at the ECB rate of 1.1205 USD per euro on 22 October 2026.
See the full calculation →
