Why it matters
Without it, misdescriptions, inconsistent quotes, delayed replies, and lost opportunities occur as questions go unanswered or poorly answered, causing unsold stock and damaged client trust.
From design to safety
Design
AI-ready enquiry standards defined
Define data schemas, catalogue standards, and response templates. AI reads inquiries, condition questions, and photo requests to shape consistent, compliant replies. Data schemas include ESTC points, imprint history, binding notes, provenance cues, and copyright constraints; templates cover initial acknowledgement, interim quotes, and follow-up prompts.
- quotation ledger
- condition report form
- provenance certificate
Build
Ingest and validate enquiry inputs
Integrate inquiry forms, photo uploads, and catalog data; establish validation, alignment with ESTC/first edition notes, and triggers for follow-up. Validation checks include item ID concordance, edition status cross-check, language in questions, and file-size thresholds for photos.
- customer inquiry log
- photograph dossier
- sales catalog entry
Operation
Respond and quote with AI-assisted drafts
AI drafts replies, checks pricing against catalog data, generates quotes, requests extra photos, and schedules follow-ups; human reviews final text. Follow-ups include estimated dispatch dates, shipping methods, and insurance options; quotes embed edition points and provenance notes.
- quotation ledger
- customer inquiry log
- packing slip
Maintenance
Audit responses and refine prompts
Regularly review responses for accuracy, update price rules, condition terms, and provenance notes; monitor data drift and photo completeness. Maintain a changelog for prompt revisions, re-train on new ESTC references, and re-validate catalog links.
- condition report form
- provenance certificate
- photograph dossier
Safety
Provenance and compliance guardrails
Ensure replies respect provenance due diligence, copyright, and sale restrictions; secure client data and prevent misrepresentation or illicit item handling. Implement two-person review for high-value items and maintain audit trails for all exchanges.
- provenance certificate
- sales catalog entry
- condition report form
Part by part
- 1
customer inquiry log
- Watch for
- Incomplete inquiries or non-standard terms that could mislead pricing; missing edition points or provenance hints.
- AI action
- AI reads inquiry intent, flags for human quote, and schedules a follow-up task.
- 2
condition report form
- Watch for
- Discrepancies between stated condition and observed faults; unrecorded foxing or restoration marks.
- AI action
- AI suggests precise condition wording, notes faults, and requests targeted photos.
- 3
provenance certificate
- Watch for
- Missing or dubious chain of title or ownership notes; gaps in archival sources.
- AI action
- AI flags provenance gaps, prompts for archival references, and assigns verification task.
- 4
photograph dossier
- Watch for
- Low-resolution or missing decisive photos that impede evaluation; missing close-ups of bindings.
- AI action
- AI requests higher-res images, specifies angles, and logs the brief in dossier.
- 5
sales catalog entry
- Watch for
- Inconsistent pricing, unsophisticated description, or missing bibliographic data; missing ESTC points.
- AI action
- AI drafts listing with edition notes, ESTC references, and provenance reminders.
- 6
packing slip
- Watch for
- Inaccurate item identification or shipping restrictions; fragile packaging notes.
- AI action
- AI confirms item ID, adds insurance notes, and outputs packaging instructions.
Worked example
Catching a missing photo before a quote
Illustrative scenario with sample data.
A dealer receives an inquiry about a first edition of a 1935 travel narrative; customer requests more plates and condition specifics.
Signal in
inquiry_id: Q-1123; item: 'First edition, 1935 travel narrative'; questions: ['condition of plates','any foxing','ESTC points']; photos: '2 interior shots'
System out
alert: missing plates photos; action: request high-res plate images, summarize condition notes, provisional quote pending photos; nextStep: schedule follow-up call.
Human sign-off: Cataloguer reviews and approves the photo request, then authorizes AI to draft the updated reply and quote.
How to roll it out
Phase 1
Weeks 1 to 4
Baseline setup of data schemas, templates, and access for inquiry handling.
Phase 2
Weeks 5 to 8
Integrate live pricing rules, photo upload workflows, and human review gates.
Phase 3
Weeks 9 to 12
Fine-tune prompts, establish QA cadence, and enforce provenance checks.
Where people stay in charge
- Decisions on pricing and quoting must be approved by a licensed cataloguer or appraiser before sending.
- Provenance and copyright restrictions must be verified by the responsible custodian before any listing or quote is published.
- Client data handling and storage must comply with applicable records privacy requirements and be auditable.
Questions owners and crews ask
How should inquiries be categorized for AI routing in pricing and listings?
Classify by item type, edition status, and requested information; route to the cataloguer for pricing, to the conservator for condition queries, or to client services for follow-ups.
What data is essential to generate an accurate quote?
Edition points, binding status, observed faults, special features, and provenance notes; AI combines these with current catalog pricing and ESTC references.
How are extra photos requested and tracked?
AI flags missing angles, creates a photo brief, emails the client with a link, and logs the request in the photograph dossier.
Who approves the final reply and quote?
A qualified cataloguer or appraiser reviews AI-generated drafts, verifies condition, provenance, and price integrity, then signs off before sending to the client.