Case study
AI-Assisted Quoting Without Invented Prices
AI-assisted quotation preparation. Rule-based pricing.
The challengeManual quote preparation and pricing consistency.
The solutionHistorical matching, pricing rules, and human exception review.
Prices calculated through defined rules, not invented by a model.
Human consultation when historical matches are unavailable.
An average of 70 quotes per day reported.
Engagement
Overview
MavenIT needed a more structured way to prepare equipment-support quotations. The workflow had to use historical knowledge while keeping pricing tied to equipment, contract information, and defined business rules.
I built a guided quote-generation workflow that retrieves relevant historical quotes and combines them with structured inputs. The language model supports interpretation and drafting, while deterministic rules govern pricing; cases without a suitable historical match go to a person for review.
Reported usage includes 22 users and an average of 70 quotes per day. These are internally reported usage figures; the reporting period and definition of a user remain to be confirmed.
- Reported Users
- 22
- Reported quotes / day, average
- 70
Challenge
Automating quote preparation was only useful if the workflow preserved control over pricing.
Historical context
Relevant past quotations needed to be matched to the current equipment and contract request.
Pricing consistency
Calculated prices needed to follow defined rules rather than model-generated estimates.
Exception handling
Missing or unsuitable historical matches needed a human decision, not a fabricated substitute.
Controlled access
The quote workflow needed access and audit controls alongside generation.
Solution
I separated language-model assistance from the rules that determine the final price.
Guided inputs
Capture equipment and contract information through a quote workspace.
Historical matching
Retrieve relevant quote records to inform the current request.
Deterministic pricing
Apply defined pricing rules to matched equipment, contract terms, and reviewed inputs.
Human exception review
Route cases without a suitable match for consultation before pricing proceeds.
Access and audit controls
Include authorization and audit events in the workflow.
Impact
Reported daily quote volume shows recurring operational use, with explicit controls around pricing and exceptions.
Adoption
22 reported users.
Quote volume
An average of 70 quotes generated per day reported.
Pricing control
Defined rules govern calculation, with human consultation when historical matching is insufficient.