Predictika · Ion implanter quotes

An AI sales agent's ion implanter quote, checked before it goes out

An AI sales agent prepares a quote for a Purion ion implanter: the model and its application class, the wafer size, the substrate, the implant temperature, the implant energy, the number of tools, the shipping region, the discount and the net total the quote states. Before it goes out, the agent sends it to Predictika, which checks it against a declarative model of how those choices fit together and returns a verdict, every reason and the totals it computed.

Scope: One ion implanter quote: Purion model, wafer size, substrate, implant temperature and energy, tools, region, discount and the stated net total. See everything this validator covers

This validator models Purion implanter quotes, not just the scenarios shown below.

  • Model and application — Purion H (high current), M (medium current), XE and XEmax (high energy) and the Purion Power Series+ models H200+, M+, XE+ and EXE+, each for its own application class and implant energy range.
  • Wafers, substrates and temperature — Power Series+ models take 150 and 200 mm wafers, silicon carbide, GaN and GaAs substrates and warm or hot implants; the other Purion models are modeled for 300 mm silicon at room temperature, and a hot implant is limited to silicon and silicon carbide.
  • Quantity, region and discount — an order needs at least one tool, more than three tools, a shipment to China or a discount over 8 percent goes to a person, and a discount over 15 percent is blocked.
  • Totals — the list total and the net total are computed from the tools, the hot or warm implant option and the discount, and the net total the quote states is compared with the computed one.

Any Purion quote that touches these rules is in scope — not only the scenarios pictured here. Not covered: other Axcelis product lines, service, spares and upgrades. Model names, application classes and the Power Series+ wafer sizes, substrates and temperatures follow public Axcelis pages; all prices, energy ranges and approval thresholds are modeled from comparable products and are illustrative, not published by Axcelis.

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

6 scenarios in 2 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
A model that fits only certain wafers, substrates and temperaturesOnly with care
Each model has its own wafer sizes, substrates and temperatures, so every model needs its own rules against each choice, and a new model means a new set of rules.
The model is one choice and every wafer, substrate and temperature is checked against it in the same pass.
An implant energy range for each modelOnly with care
The energy has to sit inside the range of the chosen model, so each model carries its own lower and upper limit and a changed model can silently invalidate the energy.
The energy is checked against the range of the model on every call.
A hot implant that depends on substrate and applicationOnly with workarounds
A hot implant is allowed on silicon and silicon carbide but not on GaN or GaAs, and on a high-energy tool it needs a person to confirm, so three choices have to be read together.
The temperature, the substrate and the application are checked together in one rule set.
A discount and a net total that must match the computed onesOnly with workarounds
The list total depends on tools and options, the discount limits depend on that total, and a stated net total has to be recomputed and compared.
The totals are computed inside the model and every threshold is checked against them.
Every problem returned in one answerOnly with workarounds
Rules tend to fire as a quote is saved or a step is completed, so a rep often fixes one error and meets the next.
All findings come back together, so the agent can fix them in a single pass.

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 needs an implant energy of 4,000 keV; this model reaches 3,000 keV.
2The quote needs an implant energy of 5 keV; this model starts at 10 keV.
3The application on the quote does not match the model; H models are high current, M models medium current, XE, XEmax and EXE models high energy.
4The quote has 200 mm wafers; Purion Power Series+ models take 150 mm and 200 mm wafers, the other Purion models are modeled for 300 mm.
5SiC, GaN and GaAs substrates are supported on the Purion Power Series+ models; this model is modeled for silicon only.
6Warm and hot implants are offered on the Purion Power Series+ models; this model is modeled for room-temperature implants.
7A hot implant is modeled for silicon and silicon carbide only, not GaN or GaAs.
8The quote asks for a hot implant on a high-energy tool; a person should confirm the process window with applications engineering.
9The quote has 0 tools; an order needs at least 1.
10The quote has 4 tools; orders of more than 3 tools go to a person for delivery-slot confirmation.
11The tool ships to China; a person must complete the export-control review before the quote is released.
12The quote gives a discount of 601 thousand USD on a list total of 7,500 thousand USD, which is over 8 percent and needs sales management approval.
13The quote gives a discount of 1,126 thousand USD on a list total of 7,500 thousand USD, which is over 15 percent.
14The quote states a net total of 7,500 thousand USD; list price plus options less discount gives 6,900 thousand USD.

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. Purion M quote for a fab: A Purion M quote with 200 mm silicon carbide wafers, a hot implant at 900 keV, four tools to China, a large discount and a wrong total

The agent's request. Please quote 4 Purion M tools for medium current implants on 200 mm silicon carbide wafers with a hot implant at 900 keV, shipping to China. The quote gives a discount of 5,000 thousand USD. It states a net total of 27,000 thousand USD and is ready to send.

Verdict: Blocked.

  • Wafer size not supported by this model. The quote has 200 mm wafers; Purion Power Series+ models take 150 mm and 200 mm wafers, the other Purion models are modeled for 300 mm.
  • Substrate not supported by this model. SiC, GaN and GaAs substrates are supported on the Purion Power Series+ models; this model is modeled for silicon only.
  • Warm or hot implant not supported. Warm and hot implants are offered on the Purion Power Series+ models; this model is modeled for room-temperature implants.
  • Energy above the model range. The quote needs an implant energy of 900 keV; this model reaches 750 keV.
  • Discount over 15 percent. The quote gives a discount of 5,000 thousand USD on a list total of 32,400 thousand USD, which is over 15 percent.
  • Stated net total does not match. The quote states a net total of 27,000 thousand USD; list price plus options less discount gives 27,400 thousand USD.
  • Shipment to China needs export review. The tool ships to China; a person must complete the export-control review before the quote is released.
  • More than 3 tools on one order. The quote has 4 tools; orders of more than 3 tools go to a person for delivery-slot confirmation.

2. Purion M quote for a fab: The agent resubmits after fixing the errors

The agent's request. Please quote 4 Purion H200+ tools for high current implants on 200 mm silicon carbide wafers with a hot implant at 900 keV, shipping to China. The quote gives a discount of 4,000 thousand USD. It states a net total of 26,400 thousand USD and is ready to send.

Verdict: Needs approval.

  • Shipment to China needs export review. The tool ships to China; a person must complete the export-control review before the quote is released.
  • More than 3 tools on one order. The quote has 4 tools; orders of more than 3 tools go to a person for delivery-slot confirmation.
  • Discount over 8 percent. The quote gives a discount of 4,000 thousand USD on a list total of 30,400 thousand USD, which is over 8 percent and needs sales management approval.
  • Quote summary. 4 tool(s), list total 30,400 thousand USD, net total 26,400 thousand USD.

3. Purion M quote for a fab: The agent resubmits after fixing the errors

The agent's request. Please quote 3 Purion H200+ tools for high current implants on 200 mm silicon carbide wafers with a hot implant at 900 keV, shipping to the United States. The quote gives a discount of 1,800 thousand USD. It states a net total of 21,000 thousand USD and is ready to send.

Verdict: Allowed.

  • Quote summary. 3 tool(s), list total 22,800 thousand USD, net total 21,000 thousand USD.

4. Purion XE+ quote for a power-device fab: A Purion XE+ quote for a hot high-energy implant on GaN wafers with a wrong total

The agent's request. Please quote 1 Purion XE+ tool for high energy implants on 150 mm GaN wafers with a hot implant at 2,500 keV, shipping to the United States. It states a net total of 9,800 thousand USD and is ready to send.

Verdict: Blocked.

  • Hot implant on GaN or GaAs. A hot implant is modeled for silicon and silicon carbide only, not GaN or GaAs.
  • Stated net total does not match. The quote states a net total of 9,800 thousand USD; list price plus options less discount gives 10,100 thousand USD.
  • Hot high-energy implant to confirm. The quote asks for a hot implant on a high-energy tool; a person should confirm the process window with applications engineering.

5. Purion XE+ quote for a power-device fab: The agent resubmits after fixing the errors

The agent's request. Please quote 1 Purion XE+ tool for high energy implants on 150 mm silicon carbide wafers with a hot implant at 2,500 keV, shipping to the United States. It states a net total of 10,100 thousand USD and is ready to send.

Verdict: Needs approval.

  • Hot high-energy implant to confirm. The quote asks for a hot implant on a high-energy tool; a person should confirm the process window with applications engineering.
  • Quote summary. 1 tool(s), list total 10,100 thousand USD, net total 10,100 thousand USD.

6. Purion XE+ quote for a power-device fab: The agent resubmits after fixing the errors

The agent's request. Please quote 1 Purion XE+ tool for high energy implants on 150 mm silicon carbide wafers with a warm implant at 2,500 keV, shipping to the United States. It states a net total of 9,800 thousand USD and is ready to send.

Verdict: Allowed.

  • Quote summary. 1 tool(s), list total 9,800 thousand USD, net total 9,800 thousand USD.

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 a Purion ion implanter: the model and its application class, the wafer size, the substrate, and more. Before it goes out, the agent sends it to Predictika, which checks it against a declarative model of how those choices fit together and returns a verdict, every reason and the totals it computed.
  2. I will start with the first request by the agent.
  3. The agent says. Please quote 4 Purion M tools for medium current implants on 200 mm silicon carbide wafers with a hot implant at 900 keV, shipping to China. The quote gives a discount of 5,000 thousand USD. The rest of the request is on screen. It states a net total of 27,000 thousand USD and is ready to send.
  4. Before it goes out, the agent calls the validator.
  5. Blocked, with six findings and two approvals, all in one answer. The quote has 200 mm wafers; Purion Power Series+ models take 150 mm and 200 mm wafers, the other Purion models are modeled for 300 mm. SiC, GaN and GaAs substrates are supported on the Purion Power Series+ models; this model is modeled for silicon only. Warm and hot implants are offered on the Purion Power Series+ models; this model is modeled for room-temperature implants. The other three are on screen.
  6. The validator returned six blocking findings and two approvals 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 discount and a net total that must match the computed ones, a rules-based CPQ can do this only with workarounds.
  8. Now the agent resubmits.
  9. The corrected request. It changes the model to Purion H200+, the application to high current, the discount to 4,000 and the net total stated on the quote to 26,400.
  10. It calls the validator again.
  11. Not blocked, but held for a person.
  12. The six blocking findings are cleared; three approvals remain for a person: shipment to China needs export review, more than 3 tools on one order and discount over 8 percent.
  13. The agent resubmits once more.
  14. The revised request. It changes the tools to 3, the shipping region to United States, the discount to 1,800 and the net total stated on the quote to 21,000.
  15. It calls the validator once more.
  16. This time the validator clears it.
  17. All three earlier findings are cleared.
  18. Eight problems caught in one call, fixed in two passes, 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.