Predictika · NetApp Keystone storage-as-a-service
An AI sales agent prepares a NetApp Keystone request: a new subscription or a change to an existing one, with the platform, service level, committed capacity per instance, burst limit, billing frequency and term. Before it goes out, the agent sends it to Predictika, which checks it against a declarative model of NetApp's published Keystone rules and returns a verdict, every reason, and the total commitment and burst limit it computed.
Source and independence. The rules come from NetApp's public Keystone documentation: the service levels, capacity requirements, supported storage capacity and pricing pages. NetApp is not involved in this demonstration and has not endorsed it. The customer is fictional. This validator produces no prices and no line-item quote: NetApp does not publish Keystone rates, so any payment amount a request carries is supplied by the customer's own system and is labeled that way.
This validator handles any new Keystone subscription or change request on these platforms — not just the scenarios shown below.
Not covered: Keystone rates and prices, true-up, renewal terms and Keystone Sovereign, because NetApp does not publish them; hardware purchases; and NetApp's own order system. Inputs supplied by the customer rather than published by NetApp: the expected consumption, the annual payment after a reduction, the days left to expiry and the reductions already used. The cases here are examples; the rules are NetApp's.
7 scenarios in 3 groups · each group runs from the agent’s first attempt to the outcome · click one to switch
Illustrative agent statement. The validator receives the structured request below.
Proposing agent: AI sales agent (scripted)
Before sending, the agent must call Recorded run
What this means
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 caught | Typical rules-based CPQ | Predictika |
|---|---|---|
| Minimum and 25 TiB step by platform and service level, with a different unit per platform (service level instance, AFX cluster, order) | Only with care A product rule per platform. It works, but each platform adds a minimum, a step and a unit that must be kept in step with NetApp's tables. | One platform table; the checks follow from the platform chosen, in any order. |
| A reduction that must satisfy seven conditions together: Annual in Advance, term of at least 24 months, 60 days' notice, the number allowed by term, 25 percent cap, platform minimum and the $200,000 payment floor | Only with workarounds Several of the inputs, such as notice given and reductions already used, are not on the quote and have to be added as fields, and the rules interact. Usually built as separate validation rules that report one at a time. | All seven are checked on every call and every failure comes back together. |
| Increase blocked in the last 90 days before expiry unless renewing | Only with workarounds Days to expiry live on the contract, not the quote, so the rule needs the contract date copied onto the quote. | The days left and the renewal flag travel with the request and are checked with everything else. |
| Total commitment and burst limit stated in the request vs the ones the rules produce | Only with care A validation rule can compare a stated figure with a calculated one, but only if the stated figure is a field on the quote, which it rarely is. | The stated figures travel with the request and are compared with the computed ones on every call. |
| Every problem returned in one answer | Only with workarounds Rules tend to fire as the quote is saved or 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. |
Every rule is checked on every call, in any order the choices arrive; one problem never hides another. Blocking findings stop the request; warnings and notes are returned with it.
| 1 | The Keystone Value performance service level is not offered on AFF A-Series; AFF A-Series offers the Extreme and Premium levels. |
| 2 | A committed capacity of 150 TiB per AFX cluster is below the 200 TiB minimum for AFX. |
| 3 | A committed capacity of 150 TiB per AFX cluster is not a multiple of the 100 TiB increment for AFX. |
| 4 | AFX capacity is committed and metered as physical capacity; logical capacity was selected. |
| 5 | A burst limit of 40 percent is available only for unified and block-optimized storage; AFX offers only the default 20 percent. |
| 6 | A committed capacity of 550 TiB per order is not a multiple of the 100 TiB increment for StorageGRID SG6160. |
| 7 | A committed capacity of 3 TiB per order is below the 4 TiB minimum for Cloud Volumes ONTAP. |
| 8 | Cloud Volumes ONTAP capacity is committed and metered as logical capacity; physical capacity was selected. |
| 9 | A burst limit of 50 percent is not offered; the choices are 20, 40 or 60 percent. |
| 10 | Committed capacity cannot be increased in the last 90 days before the subscription expires unless it is being renewed; 60 days left. |
| 11 | An increase must raise the commitment; 100 TiB per instance is not above the current 100 TiB. |
| 12 | Committed capacity for cloud services cannot be reduced during the subscription term. |
| 13 | A reduction applies only to subscriptions on the Annual in Advance payment model; this subscription is billed monthly. |
| 14 | A reduction needs a subscription term of at least 24 months; this term is 12 months. |
| 15 | A 36-month term permits 1 reduction in total; 1 already used. |
| 16 | A reduction cannot take the committed capacity to 45 TiB per service level instance, below the 50 TiB minimum. |
| 17 | Annual payments after the reduction must stay at or above $200,000; the payment stated is $150,000. |
| 18 | The Keystone Premium performance service level is not offered on FAS; FAS offers the Value level. |
| 19 | Moving to a lower performance service level is not permitted; Premium is below Extreme. |
| 20 | A reduction of 35 percent exceeds the 25 percent limit per request. |
| 21 | A reduction request must reach NetApp in writing at least 60 days before the next annual period; 45 days of notice was given. |
| 22 | The request states 240 TiB committed in total; the validator computes 250 TiB. |
| 23 | The request states a burst limit of 288 TiB; the validator computes 300 TiB. |
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.
The agent's request. Please reduce the Extreme subscription on AFF A-Series from 60 to 40 TiB per service level instance (1 instance), billed annually in advance on a 36-month term, with 45 days of written notice. Annual payments after the reduction will be $150,000. The customer expects to use about 44 TiB. The request states 40 TiB total committed and a burst limit of 48 TiB. It is ready to send.
Verdict: Blocked.
The agent's request. Please reduce the Extreme subscription on AFF A-Series from 60 to 50 TiB per service level instance (1 instance), billed annually in advance on a 36-month term, with 45 days of written notice. Annual payments after the reduction will be $260,000. The customer expects to use about 44 TiB. The request states 50 TiB total committed and a burst limit of 60 TiB. It is ready to send.
Verdict: Blocked.
The agent's request. Please reduce the Extreme subscription on AFF A-Series from 60 to 50 TiB per service level instance (1 instance), billed annually in advance on a 36-month term, with 75 days of written notice. Annual payments after the reduction will be $260,000. The customer expects to use about 44 TiB. The request states 50 TiB total committed and a burst limit of 60 TiB. It is ready to send.
Verdict: Allowed.
The agent's request. Please set up a Keystone Standard subscription on AFF C-Series: 2 service level instances at 75 TiB each, committed on logical capacity, 20 percent burst limit, billed annually, 36-month term. The customer expects to use about 230 TiB. The request states 150 TiB total committed and a burst limit of 180 TiB. It is ready to send.
Verdict: Blocked.
The agent's request. Please set up a Keystone Standard subscription on AFF C-Series: 2 service level instances at 125 TiB each, committed on logical capacity, 20 percent burst limit, billed annually, 36-month term. The customer expects to use about 230 TiB. The request states 250 TiB total committed and a burst limit of 300 TiB. It is ready to send.
Verdict: Allowed.
The agent's request. Please set up a Keystone Premium subscription on ASA A-Series: 3 service level instances at 130 TiB each, committed on logical capacity, 20 percent burst limit, billed annually, 36-month term. The customer expects to use about 340 TiB. The request states 390 TiB total committed and a burst limit of 468 TiB. It is ready to send.
Verdict: Blocked.
The agent's request. Please set up a Keystone Premium subscription on ASA A-Series: 3 service level instances at 125 TiB each, committed on logical capacity, 20 percent burst limit, billed annually, 36-month term. The customer expects to use about 340 TiB. The request states 375 TiB total committed and a burst limit of 450 TiB. It is ready to send.
Verdict: Allowed.
The text of the narrated walkthrough, step by step.