Skip to content

RAG chatbot — short email questionnaire (10 questions)

A condensed version of the full client questionnaire designed to be pasted into the body of an email rather than sent as an attached document. Use it when:

  • The client is at the curious-but-not-committed stage and a 40-question document would feel heavy.
  • We need a quick, defensible cost band before the first call.
  • The contact is a busy decision-maker who will reply on their phone.

The 10 questions below are the highest-signal subset — they map to the biggest pricing variables in the quote guide. The remaining questions can be picked up on the discovery call.

How to use

  1. Copy the Email body block below into a new email.
  2. Replace [first name] and the signoff details.
  3. Send. Expected reply length: 5-10 minutes of the client's time.
  4. When the reply arrives, copy the answers into a per-client YAML and run the full pack generator for the proper proposal.

Suggested subject lines

  • Quick scoping questions — AI chatbot for your internal documents
  • 10 questions so we can put a real number in front of you
  • Discovery — AI chatbot for [Client] (10 quick questions)

Email body — copy from here

Hi [first name],

Thank you for the conversation about an AI chatbot for your internal documents. Rather than send a long form, here are the ten questions that most affect the shape and cost of a project like this. Feel free to answer in single sentences — anything you are not sure about, leave blank and we will pick it up on the call.

1. In one sentence, what should the chatbot do, and for which staff group? (e.g. "Answer HR policy questions for all employees".)

2. Roughly how many documents are in scope, and where do they live today? (A bucket is fine: under 1,000 / 1,000-5,000 / 5,000-25,000 / more. Locations such as SharePoint, shared drive, intranet, mixed.)

3. What formats are these documents in? (Word, PDF, scanned PDF, PowerPoint, web pages, other.)

4. How many staff in total would have access, and how many would you expect to use it on a typical day?

5. Do all users see the same documents, or do some documents belong only to certain roles? (e.g. HR-only material, manager-only material.)

6. Which system manages your staff logins today? (Microsoft Entra / Azure AD, Google Workspace, Okta, other.)

7. Where would you like staff to use the chatbot? (Inside Microsoft Teams, embedded on your intranet, a dedicated internal website, or no preference yet.)

8. Which best describes your position on AI data handling?

  • A. Comfortable with leading commercial AI providers (Anthropic, OpenAI, Google) under enterprise terms with UK / EU hosting.
  • B. Prefer the AI model to run in our own cloud tenant or a controlled region.
  • C. No data may leave our premises — needs to run on hardware we own.
  • D. Not sure yet, would like to discuss the trade-offs.

9. Are there compliance or regulatory regimes we should design for from day one? (GDPR / DPIA, ISO 27001, SOC 2, sector-specific such as PCI-DSS / HIPAA / FCA, etc.)

10. Is there a target go-live date, and an approved budget range for either the build or the monthly run cost? (A range is fine — it helps us tailor the proposal rather than guess.)

Once we have your replies we will send back a short written summary of what we heard, followed by a proposal with a clear from-to range for both the one-off build and the steady-state monthly cost.

Many thanks,

[Your name] Eidos Global [email] · [phone]

Email body — copy to here

Mapping back to the full questionnaire

For the internal team — each short-form question lines up with one or more questions in the full questionnaire:

Short # Covers full-questionnaire Q# Drives
1 1, 2 Scope and audience
2 6, 7 Corpus size and ingestion effort
3 8 Parser effort, OCR yes/no
4 14, 15 Token spend envelope
5 17 Permission-aware retrieval (single biggest hidden cost)
6 18 SSO integration effort
7 21 UI / integration surface
8 24 LLM choice — biggest cost variable on the quote
9 26 Guardrails and evidence-pack effort
10 36 + budget Qs Timeline and budget anchoring

If the client gives a thorough reply to the 10 short questions, that is usually enough to produce a quote band (not a fixed price). The full questionnaire becomes the discovery-call checklist rather than a form to mail.