AI CPQ Solver
Customer entered bay width and duty cycle. CPQ returned forty-two valid configs. The rep still escalated to engineering. Mercura AI solver ranks rule-valid options against stated goals with reasoning your team can override.
AI solver · runtime
Requirements → rules first → ranked pick
1 · Customer requirements
2 · Rules engine runs first
Only valid configs continueVALID
Enter the search
BLOCKED BY RULE
Trolley T4 + rail R2
3 · Goal weighting
4 · Ranked recommendation · rep can override
Scores are illustrative · every row already passed the rules engine
What the solver does
Best-fit ranking
Valid options scored against the goals you configure
Rules-grounded
Only rule-valid configurations enter the search
Human-reviewable
Reps review the reasoning and can override before quote send
The challenge
Customer entered bay width and duty cycle. CPQ returned forty-two valid configs. The rep still escalated to engineering.
A builder of overhead bridge crane systems for manufacturing and logistics bays configures span, capacity class, hook height, runway speed, and control package in Mercura CPQ. Rules block incompatible trolley and rail combinations. A buyer enters bay width, required lift capacity, and expected duty cycle. CPQ returns dozens of rule-valid line items that all price correctly.
Inside sales still pings application engineering because none of the forty-two options clearly wins on cost, lead time, and floor clearance together. Reps pick the third ranked config by habit. Customers receive quotes that work on paper but are heavier or slower than the application needs. Expert judgment does not scale when every inquiry leaves a wide valid set.
AI CPQ implementation bootstraps draft rules from ERP and specification uploads. Conversational CPQ parses buyer language during quoting. Agentic CPQ governs procurement agent endpoints. Rules-based configuration rejects invalid combinations but does not rank valid ones. The AI CPQ solver is different: Mercura searches the rule-valid configuration space, scores candidates against cost, performance, lead time, and regulatory objectives you define, and returns ranked recommendations with plain-language reasoning reps review before quote send.
Inquiry to config to price to approval to order should not queue behind engineering every time constraints leave more than one correct answer.
How it works
How the Mercura AI CPQ solver works
Buyers or reps enter application requirements: span, capacity, duty class, clearance limits, budget ceiling, or regulatory tags. Mercura runs the rules engine first so only valid configurations enter the solver search. The AI evaluates candidates against weighted objectives your product team configures, cost, performance, lead time, and compliance priorities. Ranked results include comparison rationale and confidence indicators. When trade-offs exist, the solver surfaces Pareto-optimal alternatives instead of hiding the tension. Sales engineers accept, adjust, or override recommendations before pricing and quote release. Historical approved configurations inform scoring over time. Mercura does not replace rules authoring, approval policy, or SME sign-off on novel applications.
Rules first
Validity comes before ranking
Illustrative flow · numbers vary by catalog
Customer requirements
Span · capacity · duty class
Rules engine
Incompatible combinations removed
Valid configurations
Dozens may remain
AI ranking
Scored against configured goals
Ranked recommendation
Rep reviews, accepts or overrides
Who decides what
Rules decide validity. AI ranks the valid set.
Rules engine
Is this configuration allowed?
Deterministic constraints from your published rules. The solver never bypasses them.
AI solver
Which valid option fits best?
Ranks rule-valid options against goals your team configures, with rationale reps can review.
Trade-offs
When goals conflict, the tension stays visible
Illustrative demo data · every point is a rule-valid configuration
Pareto-optimal alternative
VFD hoist · stock rail
No other valid option is better on both cost and lead time, so it stays on the shortlist.
Reps review and can override before quote send.
What's included
What the AI CPQ solver covers
01Requirements
- Requirement inputs mapped to rule-valid configuration search
- API access for solver calls in custom sales tools
02Score
- Multi-objective scoring across cost, performance, and lead time
- Learning signal from approved configurations and outcomes
03Rank
- Ranked recommendations with plain-language comparison rationale
- Pareto-optimal alternatives when goals conflict
- Confidence indicators for how well each option meets requirements
04Review
- Rep review and override workflow before quote send
The difference
Complex configuration before and after AI ranking
Rep picks from a wide valid set
- 01 Dozens of rule-valid configs with no clear best fit
- 02 Quality depends on which engineer or rep handles the inquiry
- 03 Complex quotes queue behind application engineering
- 04 Self-service stops when more than one valid answer exists
- 05 Customers receive workable but suboptimal specifications
With Mercura AI solver
- 01 Ranked shortlist from customer requirements in one session
- 02 Same scoring logic applied to every inquiry and channel
- 03 Engineering reviews exceptions, not every multi-option quote
- 04 Self-service extends to applications with many valid paths
- 05 Quotes reflect best-fit trade-offs reps can explain to buyers
Real-world example
Example workflow: crane config ranked by span, duty, and lead time
An overhead crane OEM saw reps escalate whenever bay width and duty cycle left more than thirty valid Mercura configurations. After enabling the AI solver, a buyer entered span, capacity class, and expected picks per hour. Mercura returned a ranked shortlist with reasoning on rail weight, motor package, and quoted lead time. The rep accepted the top recommendation, adjusted control package for a regional safety tag, and sent the quote without an engineering callback. Application engineers now handle novel bay layouts only.
Business impact
Why the AI solver is ranking intelligence, not a replacement for rules
The AI CPQ solver adds engineering judgment at quote time when rules alone leave multiple correct answers. It complements AI CPQ implementation, sales guidance flows, rules-based validation, and self-service configurators. Mercura does not bypass the constraint engine or publish recommendations without rep review. Someone must define objectives, validate scoring after catalog changes, and own approval on edge cases. If the pain is "CPQ says valid but nobody knows which valid option to quote", the solver aligns inquiry, configuration, price, approval, and order with ranked configs your team can defend to the buyer.
Business impact
Clearer best-fit choices
Rule-valid options are ranked against the goals your product team configures.
Rules stay in charge
Only valid configurations enter the search; the constraint engine is never bypassed.
Explainable and overridable
Reps review the rationale and can adjust or override before quote send.
AI lifecycle map
Two different jobs: bootstrap the CPQ model from existing data, then rank valid configurations at quote time
Build time
AI drafts. Your team approves before publish.
Existing data
ERP · specs · price book
AI draft
AI CPQ Implementation
Human review
Ready for review
Published rules
Rules engine
Runtime
Rules decide validity. AI ranks what is valid.
Requirements
Customer inputs
Rules engine
Valid set first
AI ranking
AI CPQ Solver
Rep review
Accept or override
Published rules from build time become the rules engine that gates every runtime recommendation
Rules-valid set · ranked for rep review
Enter span and duty cycle and review a ranked crane shortlist before quote send
Book a demo and walk requirement input through rule-valid search, ranked recommendations, and rep override, so engineering time goes to the cases that need it.
Let’s build together.
We empower manufacturers to master product modeling, streamline quoting process, reduce errors, and ultimately deliver the tailored solutions that customers demand.