Predictika · Fuel cell power and heat capture site orders

An AI sales agent's fuel cell site order, checked before it goes out

An AI sales agent prepares an order for a site powered by 325 kW solid oxide fuel cell Energy Servers, with or without steam, chilled-water or hot-water heat capture. Before it goes out, the agent sends it to Predictika, which checks it against a declarative model of the published data sheet figures and limits and returns a verdict, every reason and the figures it computed.

Scope: Energy Server 6.5 site orders: power-only, steam, chilling and hot-water heat capture configurations. See everything this validator covers

This validator models Energy Server 6.5 site orders with the heat capture options on the published data sheets, not just the scenarios shown below.

  • Platform and outputs — each Energy Server is 325 kW net AC. For steam (5 and 15 barg) and chilled-water (7/12 °C and 18/28 °C) configurations the data sheets list nine platform sizes (4 to 128 servers, 1.3 to 41.6 MW) with steam or chilling output and footprint, on the ground or stacked; hot-water heat capture is 2,777 kg/hr per 325 kW. The quote's stated power, steam, chilling, hot-water flow and footprint are compared with these figures.
  • Inputs and site — natural gas at 12 to 18 psig, three-phase output at 480, 415 or 400 V (380 V on the heat capture sheets), 50 or 60 Hz, and a site temperature range of -20 °C to 45 °C.
  • Needs approval — fuels other than natural gas (the data sheets say to contact Bloom Energy), platform sizes that are not listed (available on request), 380 V on a power-only order, an indoor installation, and a stated footprint where none is published.

Any Energy Server 6.5 order that touches these figures and limits is in scope — not only the scenarios pictured here. Not covered: prices (Bloom Energy does not publish prices, so this validator produces no price build-up or line-item quote), earlier Energy Server generations, grid interconnection codes, emissions, electrical efficiency and any figure that is not in the data sheets.

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

7 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
Output figures that depend on platform size and layoutOnly with care
Steam, chilling and footprint come from a table by platform size, pressure or temperature pair and layout; each table is a lookup to build and keep current.		Every size, application and layout is in one request, and the figures are looked up and compared with the quote on every call.
Stated figures that must match the data sheetOnly with care
A stated figure is a free-entry field; a rule can compare it with a looked-up value, but someone has to write one for each output.		The stated power, steam, chilling, hot-water flow and footprint travel with the order and are compared on every call.
Limits on gas pressure, voltage, frequency and site temperatureOnly with workarounds
Each limit is a validation rule on its own field; they work, but they only fire when that field is saved, so errors surface one at a time.		All limits are checked together, whatever order the choices arrive in.
Choices that are possible only by arrangement with the manufacturerOnly with workarounds
Fuels, unlisted sizes and indoor sites need a human decision; a rules-based CPQ usually forces a binary allow or block, or relies on a rep remembering to ask.		These come back as a separate "needs approval" finding, distinct from a block.
Every problem returned in one answerOnly with workarounds
Rules tend to fire as the order is saved or a step is completed, so a rep often fixes one error and meets the next.		All findings come back together.

  

  

    
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 no Energy Server; at least one 325 kW Energy Server is needed.
2The natural gas input pressure on the quote is 20 psig; the Energy Server data sheet requires 12 to 18 psig (15 psig nominal).
3The natural gas input pressure on the quote is 10 psig; the Energy Server data sheet requires 12 to 18 psig (15 psig nominal).
4The site design minimum temperature is -25 C; the Energy Server temperature range is -20 C to 45 C.
5The site design maximum temperature is 48 C; the Energy Server temperature range is -20 C to 45 C.
6The data sheets list three-phase output at 480, 415, 400 V (and 380 V on the heat capture sheets); 208 V is not a listed output voltage.
7The Energy Server output frequency is 50 or 60 Hz; the quote asks for 25 Hz.
8The quote states 1,500 kW net AC, but 4 Energy Servers at 325 kW each give 1,300 kW.
9The quote states 2,200 kg/hr of steam, but the data sheet gives 1,955 kg/hr for 10 Energy Servers at the selected steam pressure.
10The quote states 1,955 kg/hr of steam, but the data sheet gives 1,683 kg/hr for 10 Energy Servers at the selected steam pressure.
11The quote states 14,000 kWth of chilling, but the data sheet gives 14,627 kWth for 64 Energy Servers at the selected chilled-water temperatures.
12The quote states 50,000 kg/hr of hot water, but the data sheet gives 55,540 kg/hr for 20 Energy Servers.
13The quote states a heat output of 782, but a power-only configuration has no heat capture.
14The quote states a footprint of 2,211 square meters, but the data sheet gives 3,567 square meters for 64 Energy Servers in the selected layout.
15The data sheets list natural gas; biogas, blended hydrogen and hydrogen configurations require contacting Bloom Energy.
16The natural gas input pressure on the quote is 9 psig; the Energy Server data sheet requires 12 to 18 psig (15 psig nominal).
1712 Energy Servers is not one of the platform sizes listed for steam and chilling configurations; the data sheets say other sizes are available on request.
18The power-only data sheet lists 480, 415 and 400 V; 380 V appears only on the heat capture data sheets.
19The data sheets list outdoor as the Energy Server location; an indoor installation is not a listed configuration.
20The quote states a footprint, but the data sheets publish footprints only for the steam and chilling platform sizes; it cannot be checked.
21The natural gas input pressure on the quote is 22 psig; the Energy Server data sheet requires 12 to 18 psig (15 psig nominal).
22The site design maximum temperature is 50 C; the Energy Server temperature range is -20 C to 45 C.
23The quote states 20,000 kW net AC, but 64 Energy Servers at 325 kW each give 20,800 kW.
24The quote states a footprint of 3,000 square meters, but the data sheet gives 3,567 square meters for 64 Energy Servers in the selected layout.

  

  

    
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. Steam for a food plant: A 3.25 MW steam order with a high steam figure and a gas pressure of 21 psig

    
The agent's request. Please quote a site with 10 Bloom Energy Servers (3,250 kW net AC), heat capture for steam at 5 barg, running on natural gas at 21 psig, 480 V output at 60 Hz, installed outdoors, with site temperatures from 2 to 38 °C. Layout: on the ground. The quote states net power of 3,250 kW, steam output of 2,200 kg/hr and a footprint of 585 square meters. It is ready to send.

    
Verdict: Blocked.

    
    
      
  • Fuel pressure outside 12-18 psig. The natural gas input pressure on the quote is 21 psig; the Energy Server data sheet requires 12 to 18 psig (15 psig nominal).
    • 
      
  • Stated steam differs from the data sheet. The quote states 2,200 kg/hr of steam, but the data sheet gives 1,955 kg/hr for 10 Energy Servers at the selected steam pressure.
    • 
    

    
2. Steam for a food plant: The agent resubmits after fixing the errors

    
The agent's request. Please quote a site with 10 Bloom Energy Servers (3,250 kW net AC), heat capture for steam at 5 barg, running on natural gas at 15 psig, 480 V output at 60 Hz, installed outdoors, with site temperatures from 2 to 38 °C. Layout: on the ground. The quote states net power of 3,250 kW, steam output of 1,955 kg/hr and a footprint of 585 square meters. It is ready to send.

    
Verdict: Allowed.

    
    
      
  • Data sheet figures. 10 Energy Servers give 3,250 kW net AC.
    • 
    

    
3. Chilling for a data center: A 20.8 MW chilled-water order at 208 V with figures that differ from the data sheet

    
The agent's request. Please quote a site with 64 Bloom Energy Servers (20,800 kW net AC), heat capture for chilled water at 18/28 °C, running on natural gas at 15 psig, 208 V output at 60 Hz, installed outdoors, with site temperatures from -5 to 40 °C. Layout: stacked. The quote states net power of 20,800 kW, chilling output of 14,000 kWth and a footprint of 3,000 square meters. It is ready to send.

    
Verdict: Blocked.

    
    
      
  • Output voltage not offered. The data sheets list three-phase output at 480, 415, 400 V (and 380 V on the heat capture sheets); 208 V is not a listed output voltage.
    • 
      
  • Stated chilling differs from the data sheet. The quote states 14,000 kWth of chilling, but the data sheet gives 14,627 kWth for 64 Energy Servers at the selected chilled-water temperatures.
    • 
      
  • Stated footprint differs from the data sheet. The quote states a footprint of 3,000 square meters, but the data sheet gives 2,211 square meters for 64 Energy Servers in the selected layout.
    • 
    

    
4. Chilling for a data center: The agent resubmits after fixing the errors

    
The agent's request. Please quote a site with 64 Bloom Energy Servers (20,800 kW net AC), heat capture for chilled water at 18/28 °C, running on natural gas at 15 psig, 480 V output at 60 Hz, installed outdoors, with site temperatures from -5 to 40 °C. Layout: stacked. The quote states net power of 20,800 kW, chilling output of 14,627 kWth and a footprint of 2,211 square meters. It is ready to send.

    
Verdict: Allowed.

    
    
      
  • Data sheet figures. 64 Energy Servers give 20,800 kW net AC.
    • 
    

    
5. District hot water: A 6.5 MW hot-water order with a wrong power and hot water flow

    
The agent's request. Please quote a site with 20 Bloom Energy Servers (6,500 kW net AC), heat capture for hot water at 50/100 °C, running on natural gas at 15 psig, 400 V output at 50 Hz, installed outdoors, with site temperatures from -10 to 36 °C. The quote states net power of 6,000 kW and hot water flow of 50,000 kg/hr. It is ready to send.

    
Verdict: Blocked.

    
    
      
  • Stated power differs from the quote. The quote states 6,000 kW net AC, but 20 Energy Servers at 325 kW each give 6,500 kW.
    • 
      
  • Stated hot water differs from the data sheet. The quote states 50,000 kg/hr of hot water, but the data sheet gives 55,540 kg/hr for 20 Energy Servers.
    • 
    

    
6. District hot water: The agent resubmits after fixing the errors

    
The agent's request. Please quote a site with 20 Bloom Energy Servers (6,500 kW net AC), heat capture for hot water at 50/100 °C, running on natural gas at 15 psig, 400 V output at 50 Hz, installed outdoors, with site temperatures from -10 to 36 °C. The quote states net power of 6,500 kW and hot water flow of 55,540 kg/hr. It is ready to send.

    
Verdict: Allowed.

    
    
      
  • Data sheet figures. 20 Energy Servers give 6,500 kW net AC.
    • 
    

    
7. Cold site: A power-only order for a site that falls to -28 °C

    
The agent's request. Please quote a site with 8 Bloom Energy Servers (2,600 kW net AC), power only (no heat capture), running on natural gas at 15 psig, 480 V output at 60 Hz, installed outdoors, with site temperatures from -28 to 30 °C. The quote states net power of 2,600 kW. It is ready to send.

    
Verdict: Blocked.

    
    
      
  • Site colder than -20 C. The site design minimum temperature is -28 C; the Energy Server temperature range is -20 C to 45 C.
    • 
    

  

  

    
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 an order for a site powered by 325 kW solid oxide fuel cell Energy Servers, with or without steam, chilled-water or hot-water heat capture. Before it goes out, the agent sends it to Predictika, which checks it against a declarative model of the published data sheet figures and limits and returns a verdict, every reason and the figures it computed.
    2. 
      
  2. I will start with the first request by the agent.
    3. 
      
  3. The agent says. Please quote a site with 64 Bloom Energy Servers (20,800 kW net AC), heat capture for chilled water at 18/28 °C. Layout: stacked. The rest of the request is on screen. The quote states net power of 20,800 kW, chilling output of 14,000 kWth and a footprint of 3,000 square meters.
    4. 
      
  4. Before it goes out, the agent calls the validator.
    5. 
      
  5. Blocked, with three findings, all in one answer. The data sheets list three-phase output at 480, 415, 400 V (and 380 V on the heat capture sheets); 208 V is not a listed output voltage. The quote states 14,000 kWth of chilling, but the data sheet gives 14,627 kWth for 64 Energy Servers at the selected chilled-water temperatures. The quote states a footprint of 3,000 square meters, but the data sheet gives 2,211 square meters for 64 Energy Servers in the selected layout.
    6. 
      
  6. The validator returned three 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.
    7. 
      
  7. What this shows: for stated figures that must match the data sheet, a rules-based CPQ is error-prone for this.
    8. 
      
  8. Now the agent resubmits.
    9. 
      
  9. The corrected request. It changes the output voltage to 480 V, the stated chilling output to 14,627 kWth and the stated footprint to 2,211 square meters.
    10. 
      
  10. It calls the validator again.
    11. 
      
  11. This time the validator clears it.
    12. 
      
  12. All three earlier findings are cleared, and the validator's figure matches the quote.
    13. 
      
  13. Three 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.
    14.