Predictika · Rack server quotes

An AI sales agent's rack server quote, checked before it goes out

An AI sales agent prepares a quote for NEC Express5800 rack servers: the model, the number of servers, the processors, the DIMMs and their size, the 2.5-inch and 3.5-inch drives, the RAID controller, the GPU cards, the power supplies, the maintenance and its term, and the total memory it states. Before it goes out, the agent sends it to Predictika, which checks it against a declarative model of the configuration rules and returns a verdict, every reason and the total memory it computed.

Scope: NEC Express5800 rack servers (R120k-2M, R120j-2M, R125k-2M, R120k-1M, R120j-1M and R110k-1M): processors, memory, drives, RAID, GPU cards, power supplies, maintenance and the stated total memory. See everything this validator covers

This validator models quotes of this kind, not just the scenarios shown below.

  • Processors and memory — the processors each model takes, the DIMM slots they give, the offered DIMM sizes, the memory each model supports, and the stated total memory against the computed total.
  • Drives, RAID and GPU — how many 2.5-inch and 3.5-inch drives each model holds, no mixing of the two sizes, a RAID controller above two drives, and the GPU cards each model supports.
  • Power, maintenance and approvals — two power supplies for GPU cards and for 24x7 maintenance, a maintenance term of 3 to 5 years, and memory channels left unpopulated or an order of more than 50 servers, both for a person to confirm.

Any quote that touches these rules is in scope — not only the scenarios pictured here. The model names, the form factors, the processor counts, the maximum memory, the 3-year warranty and its extension to 5 years follow public NEC information; the DIMM slots and channels, the drive bays, the GPU limits, the power and RAID rules, the DIMM sizes and the 50-server threshold are modeled from comparable servers and are not published by NEC. No prices are modeled. Not covered: storage, software and other Express5800 series.

The AI sales agentBuilds the server quote from the conversation and says it is ready.
Predictika validatorChecks every rule at once and returns the verdict, each finding, and the computed figures.

6 scenarios in 3 groups · each group runs from the agent’s first attempt to the outcome · click one to switch

What the agent proposes

The AI sales agent says

Illustrative agent statement. The validator receives the structured request below.

Proposing agent: AI sales agent (scripted)

Before sending, the agent must call Recorded run

Predictika validator


      
Waiting for the call.

Where rules-based CPQ gets it wrong

Click to expand

Our assessment of what a typical rules-based CPQ (product rules, price rules and validation rules, as in Salesforce CPQ) does with the same cases. It is a judgment about effort and reliability, not a claim that nothing can be built.

What the validator caughtTypical rules-based CPQPredictika
Limits that differ for each modelOnly with care
Each model has its own processors, memory maximum, drive bays and GPU limit, so every limit needs a rule that reads the model.
The model and every option are checked together on every call.
Slots that follow the processorsOnly with care
The DIMM slots and memory channels depend on how many processors are quoted, so the limit must be computed before it is compared with the DIMMs.
The processors, DIMMs and DIMM size are checked together on every call.
Options that need other optionsOnly with workarounds
Drives above two need a RAID controller, GPU cards and 24x7 maintenance need two power supplies, and 2.5-inch and 3.5-inch drives cannot be mixed, so each rule reads other choices.
All related choices are fields of the quote and are checked together on every call.
A stated total that must match the computed totalOnly with workarounds
The total is the servers times the DIMMs times the DIMM size, so a separate rule must recompute it and compare it with the stated figure.
The stated total memory is a field of the quote and is compared with the computed total on every call.
Every problem returned in one answerOnly with workarounds
Rules tend to fire as a step is completed, so a rep often fixes one error and meets the next.
All findings come back together, whatever order the choices arrived in.

What this validator catches: findings from its test cases

Click to expand

Every rule is checked on every call, in any order the choices arrive; one problem never hides another. Blocking findings stop the quote; approval findings need a person to sign off; everything else is computed.

1The quote has 2 processors per server; this model supports at most 1.
2The quote has 20 DIMMs per server; the processors quoted give 16 DIMM slots.
3DIMM size 48 GB is not one of the offered sizes (16, 32, 64, 128, 256 GB).
4The quote has 8,192 GB of memory per server; this model supports at most 4,096 GB.
5The quote states 1,024 GB of memory in total; the quoted servers give 2,048 GB.
6The quote has 12 2.5-inch drives per server; this model holds at most 10.
7The quote has 4 3.5-inch drives per server; this model holds at most 0.
82.5-inch and 3.5-inch drives cannot be mixed in one server chassis.
9The quote has 6 drives per server and no RAID controller.
10GPU cards need two power supplies; the quote has 1.
11The quote has 1 GPU cards per server; this model supports at most 0.
1224x7 on-site maintenance needs two power supplies; the quote has 1.
13Maintenance can be extended to 5 years at most; the quote has 6 years.
14The standard warranty is 3 years; the quote has 2 years.
15The quote has 8 DIMMs for 16 memory channels; a person should confirm the reduced memory bandwidth is acceptable.
16An order of 60 servers is above the 50-server limit for standard handling; a person should approve it.

Scenarios and results

Click to expand

The recorded result of the validator for each scenario on this page: the request the AI agent sent, the verdict, and every finding, in the order they appear above.

1. Model, memory, GPU and power: An R110k-1M quote with two processors, 48 GB DIMMs, ten 2.5-inch drives, a GPU card, one power supply and 24x7 maintenance

The agent's request. Please quote 2 NEC Express5800 R110k-1M servers, each with 2 processors, 32 DIMMs of 48 GB, 10 2.5-inch drives, a RAID controller, 1 GPU card, 1 power supply and 24x7 on-site maintenance for 3 years. The quote states 3,000 GB of memory in total. It is ready to send.

Verdict: Blocked.

  • Too many processors. The quote has 2 processors per server; this model supports at most 1.
  • DIMM size not offered. DIMM size 48 GB is not one of the offered sizes (16, 32, 64, 128, 256 GB).
  • Stated memory total does not match. The quote states 3,000 GB of memory in total; the quoted servers give 3,072 GB.
  • Too many 2.5-inch drives. The quote has 10 2.5-inch drives per server; this model holds at most 8.
  • Too many GPU cards. The quote has 1 GPU cards per server; this model supports at most 0.
  • GPU needs redundant power. GPU cards need two power supplies; the quote has 1.
  • 24x7 service needs redundant power. 24x7 on-site maintenance needs two power supplies; the quote has 1.

2. Model, memory, GPU and power: The agent resubmits after fixing the errors

The agent's request. Please quote 2 NEC Express5800 R120k-2M servers, each with 2 processors, 32 DIMMs of 64 GB, 10 2.5-inch drives, a RAID controller, 1 GPU card, 2 power supplies and 24x7 on-site maintenance for 3 years. The quote states 4,096 GB of memory in total. It is ready to send.

Verdict: Allowed.

  • Server summary. 2 server(s) with 2,048 GB of memory each, 4,096 GB in total.

3. Drives and RAID: An R120k-1M quote with twelve 2.5-inch drives, four 3.5-inch drives and no RAID controller

The agent's request. Please quote 2 NEC Express5800 R120k-1M servers, each with 2 processors, 16 DIMMs of 64 GB, 12 2.5-inch drives and 4 3.5-inch drives, no RAID controller, no GPU card, 2 power supplies and standard maintenance for 3 years. The quote states 2,048 GB of memory in total. It is ready to send.

Verdict: Blocked.

  • Too many 2.5-inch drives. The quote has 12 2.5-inch drives per server; this model holds at most 10.
  • Too many 3.5-inch drives. The quote has 4 3.5-inch drives per server; this model holds at most 0.
  • 2.5-inch and 3.5-inch drives mixed. 2.5-inch and 3.5-inch drives cannot be mixed in one server chassis.
  • More than 2 drives need a RAID controller. The quote has 16 drives per server and no RAID controller.

4. Drives and RAID: The agent resubmits after fixing the errors

The agent's request. Please quote 2 NEC Express5800 R120k-2M servers, each with 2 processors, 16 DIMMs of 64 GB, 8 3.5-inch drives, a RAID controller, no GPU card, 2 power supplies and standard maintenance for 3 years. The quote states 2,048 GB of memory in total. It is ready to send.

Verdict: Allowed.

  • Server summary. 2 server(s) with 1,024 GB of memory each, 2,048 GB in total.

5. Memory and term: An R120j-2M quote with 32 DIMMs of 256 GB and a six-year maintenance term

The agent's request. Please quote 1 NEC Express5800 R120j-2M server, each with 2 processors, 32 DIMMs of 256 GB, 8 2.5-inch drives, a RAID controller, no GPU card, 2 power supplies and standard maintenance for 6 years. The quote states 8,192 GB of memory in total. It is ready to send.

Verdict: Blocked.

  • Memory over the model maximum. The quote has 8,192 GB of memory per server; this model supports at most 4,096 GB.
  • Support term over 5 years. Maintenance can be extended to 5 years at most; the quote has 6 years.

6. Memory and term: The agent resubmits after fixing the errors

The agent's request. Please quote 1 NEC Express5800 R120j-2M server, each with 2 processors, 32 DIMMs of 128 GB, 8 2.5-inch drives, a RAID controller, no GPU card, 2 power supplies and standard maintenance for 5 years. The quote states 4,096 GB of memory in total. It is ready to send.

Verdict: Allowed.

  • Server summary. 1 server(s) with 4,096 GB of memory each, 4,096 GB in total.

Narration transcript

Click to expand

The text of the narrated walkthrough, step by step.

  1. Welcome to a demo of Predictika's Validator. An AI sales agent prepares a quote for NEC Express5800 rack servers: the model, the number of servers, the processors, and more. Before it goes out, the agent sends it to Predictika, which checks it against a declarative model of the configuration rules and returns a verdict, every reason and the total memory it computed.
  2. I will start with the first request by the agent.
  3. The agent says. Please quote 2 NEC Express5800 R110k-1M servers, each with 2 processors, 32 DIMMs of 48 GB, 10 2.5-inch drives, a RAID controller. The quote states 3,000 GB of memory in total. The rest of the request is on screen.
  4. Before it goes out, the agent calls the validator.
  5. Blocked, with seven findings, all in one answer. The quote has 2 processors per server; this model supports at most 1. DIMM size 48 GB is not one of the offered sizes, such as the 16 and 32. The quote states 3,000 GB of memory in total; the quoted servers give 3,072 GB. The other four are on screen.
  6. The validator returned seven blocking findings in one answer. A check that reads one field at a time would pass most of this; it shows only when the choices in the quote are checked against each other and against the stated total.
  7. What this shows: for a stated total that must match the computed total, a rules-based CPQ can do this only with workarounds.
  8. Now the agent resubmits.
  9. The corrected request. It changes the model to R120k-2M, the DIMM size to 64 GB, the power supplies to 2 and the stated total memory to 4,096.
  10. It calls the validator again.
  11. This time the validator clears it.
  12. All seven earlier findings are cleared.
  13. Seven problems caught in one call, fixed in one pass, and a clean quote. That is the value of Predictika's validator: catching mistakes before they reach the customer. Now you can try the other scenarios yourself.