Blog · Published · Updated
Can employees use the assistant’s answer?
Short answer
You test an AI knowledge base on employees' questions and on what they do after reading the answer. An accurate link to the instruction is not enough if a new person still does not know where to get the data for the first field. An AI application on company documents costs from €3,000 net at Syntalith. At 300 employee questions a month and 12 minutes of searching each, it pays back in 1.7 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
Testing a company AI assistant: statistics 2026
1.7 mo.
payback of €3,000 at 300 employee questions a month: €2,094 of work today, €319 with checking and the model. Scenario.
15%
According to Brynjolfsson, Li and Raymond (2025), customer support agents with AI help resolved 15% more issues an hour on average, in a study of 5,172 people.
Source: Brynjolfsson, Li and Raymond, The Quarterly Journal of Economics (2025) ↗
30%
According to Brynjolfsson, Li and Raymond (2025), less experienced support agents with AI raised the number of issues resolved an hour by 30%.
Source: Brynjolfsson, Li and Raymond, The Quarterly Journal of Economics (2025) ↗
49%
According to Anthropic (2024), combining contextual embeddings with BM25 search cut failed retrievals by 49%, and by 67% with reranking.
48%
According to the European Commission (2026), 48% of small and medium-sized companies in the EU name a lack of expertise as a barrier to AI.
Source: European Commission, Flash Eurobarometer 578 (2026) ↗
Figures checked: 22 October 2026
The same instruction, a different starting point
The assistant points to the form for ordering event materials. An experienced coordinator knows which category to choose and where to get the campaign identifier. A new person sees the correct link, but stops at the first field and asks a colleague where to get the number. In the demo everything looked fine, because the author of the instruction asked the questions.
- Questions from people who do the work regularly and from new staff
- Observation up to a checked form
- A note on „what happened next” as well as a satisfaction score
- A scope that matches the planned rollout
The authors of instructions understand shortcuts and dependencies the rest of the team cannot see, so their part is needed to judge the content, while the users' experience has to come from the users. A note such as „I had to ask a colleague where to get the number” gives a concrete direction for improvement.
What it costs and when it pays back
- Today
- 300 employee questions × 12 min = 60 h (€2,094)
- With the system, a person checks the result in 15% of the time
- €314
- Model, 1 call per case, Claude Sonnet 5.5
- €5
- Cost of one case
- €6.98 today, €1.06 with the system
- Saving
- €1,775 a month
Payback of €3,000
1.7 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: 6,000 input tokens and 500 output tokens per call at Anthropic's rates, 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 remains for checking the answer. According to Brynjolfsson, Li and Raymond (2025), less experienced support agents with AI raised the number of issues resolved an hour by 30%, so an assessment with new employees shows the largest part of the effect.
Before we price the build, we run a test on 20 to 100 questions the team has already asked: we compare the application's answer with what the employee found by hand, and we count the time from the question to the source. The test takes 5 working days and costs €0.
A usage report does not replace a conversation
Copilot Studio lets you review users' questions, sessions and reactions, with access depending on the role. Such data helps choose the questions to discuss, but it will not show that an employee abandoned a task or did it the old way. The company decides who sees the notes and how long they are kept. The aim is better help with the task, and ranking people by the number of questions makes no sense here.
Seeing what people ask about
the platform's usage report
within the licence
Checking whether the answer helps with the task
watching a few people at work
part of the trial
Deciding on a wider rollout
an assessment on 20 to 100 questions with timing
test €0
Where to start
We limit the assessment to one kind of work, for example preparing requests. We invite the people who will use the tool, alongside the authors of the instructions.
For the first call, one question that keeps coming back to colleagues and where the team looks for the answer are enough. You do not need to send confidential documents, and you receive the 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 review.
€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 →
