Predictika · Software subscription quotes

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

An AI agent assembles a Loopio subscription quote (a plan, standard, library and content contributor users, add-ons, a term and discounts) and states its own totals. Before the quote goes to the customer, the agent sends it to Predictika’s validator, which checks it against the plan packaging and an example quote policy, recomputes the agent’s figures, and returns a verdict (allow, needs sign-off or block) with every reason at once. The same check works on a quote assembled by a person.

Scope: Loopio subscription quotes on four plans (Essentials, Plus, Enhanced, Enterprise): user types, add-ons, term, discounts, renewal uplift and the totals the agent states. See everything this validator covers

This validator models a Loopio subscription quote against the published plan packaging and an example quote policy, not just the scenarios shown below.

  • Plan packaging — each plan's standard user range, library users and content contributors (Plus and above), the Go/No-Go module (Plus and above), project translations and the Premium Customer Success Plan (Enhanced and above), and the integrations offered on Essentials.
  • Term, discounts and renewals — a term of at most three years, a multi-year discount only on multi-year terms, a combined discount limit, sign-off above the multi-year and competitive discount bands, a renewal uplift limit, and competitive discounts on new purchases only.
  • Totals the agent states — the annual total and the multi-year savings, recomputed from every line and compared with the figures the agent states; large annual contracts and Enterprise quotes are held for sign-off.

Any Loopio quote that touches these rules is in scope — not only the scenarios pictured here. Loopio does not publish a price list: the figures are published market ranges (Vendr, October 2026) and the policy limits are an example policy, both configurable. Not covered: real negotiated pricing, taxes, other Loopio products and one-time fees. The page shows the totals the validator computed but no line-item table.

The AI sales agentBuilds the subscription quote from the conversation, states its own totals 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
Totals the agent states itselfOnly with workarounds
The annual total and the multi-year savings are usually typed or generated by the agent; checking them means recomputing every line, both discounts and the term, which takes custom code.
Every line, discount and total is recomputed on each call and compared with the agent's figures.
Which plan allows which add-on or user typeOnly with care
Each eligibility (library users, translations, premium plan, integrations) is a separate product or option rule, and the plan ladder has to be kept in step in each of them.
All eligibility checks run on every quote and each reports its own finding.
Seats against the plan's rangeOnly with care
Seat caps and per-seat pricing beyond the included users are usually two different mechanisms that have to agree.
Seats, included users and the additional-seat line are computed together and compared with the plan range.
A discount outside its bandEasy
Discount approval thresholds are standard in quoting tools.
The same check, returned with every other finding in one answer.
Stacked discounts and a combined limitOnly with care
Multi-year and competitive discounts are often approved separately, so a combined limit needs its own summary rule.
Both discounts are added and checked against the combined limit on every call.
Renewal-only and new-purchase-only termsOnly with care
Uplift and competitive discounts depend on the quote type; rules that ignore it let the wrong one through.
Each is checked against the quote type, and a violation is its own finding.
Every problem returned in one answerOnly with workarounds
Rules tend to stop at the first failing check, so reasons surface one resubmission at a time.
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.

1Library users and content contributors need Plus or higher: the quote has 3 library users and 3 content contributors; Essentials does not include these user types.
2Standard users above the plan range: the quote has 12 standard users; Essentials covers up to 10.
3Stated annual total does not match: the agent stated $22,000 a year; the validator computes $25750.
4Standard users above the plan range: the quote has 11 standard users; Essentials covers up to 10.
5Combined discount above the policy limit: the multi-year and competitive discounts total 55 percent; the limit is 40 percent.
6Stated multi-year savings do not match: the agent stated $170,000 saved over the term; the validator computes $190080.
7Multi-year discount above the published band: 35 percent is above the 30 percent band; sign-off is needed.
8Multi-year discount above the published band: 33 percent is above the 30 percent band; sign-off is needed.
9Combined discount above the policy limit: the multi-year and competitive discounts total 41 percent; the limit is 40 percent.
10Competitive discount above the published band: 26 percent is above the 25 percent band; sign-off is needed.
11Project translations need Enhanced or higher: 3 language(s) across 4 project(s) are requested; Plus does not offer project translations.
12Stated annual total does not match: the agent stated $58,000 a year; the validator computes $59920.
13Premium Customer Success Plan needs Enhanced or higher: the plan is not offered with Plus.
14Integration not offered on this plan: the hubspot integration is not offered with Essentials; only the Salesforce add-on is.
15Renewal uplift above the policy limit: the renewal uplift is 6 percent; the limit is 5 percent.
16Enterprise price is set individually: Enterprise is quoted case by case; the deal desk confirms the price.
17Large annual contract: the annual net is $197,000; sign-off is needed above $150000.
18Term longer than the policy allows: the quote is for 4 years; the longest term is 3 years.
19Stated annual total does not match: the agent stated $37,551 a year; the validator computes $37500.

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. Plan packaging, term, discounts and totals: An Essentials quote that asks for far more than the plan offers, with ten problems at once

The agent's request. Please quote a new Essentials subscription for 12 standard users, 4-year term, with 3 library users, 3 content contributors, the Go/No-Go module, the Premium Customer Success Plan, project translations (2 languages, 3 projects), the HubSpot integration, a 35 percent multi-year discount and a 20 percent competitive discount, for $15,000 a year and $90,000 saved over the term. It is ready to send.

Verdict: Blocked.

  • Library users and contributors need Plus or higher. Library users and content contributors need Plus or higher: the quote has 3 library users and 3 content contributors; Essentials does not include these user types.
  • Standard users above the plan range. Standard users above the plan range: the quote has 12 standard users; Essentials covers up to 10.
  • Go/No-Go module not offered on this plan. Go/No-Go module not offered on this plan: the Go/No-Go Evaluation module is not offered with Essentials.
  • Project translations need Enhanced or higher. Project translations need Enhanced or higher: 2 language(s) across 3 project(s) are requested; Essentials does not offer project translations.
  • Integration not offered on this plan. Integration not offered on this plan: the hubspot integration is not offered with Essentials; only the Salesforce add-on is.
  • Premium Customer Success Plan needs Enhanced or higher. Premium Customer Success Plan needs Enhanced or higher: the plan is not offered with Essentials.
  • Term longer than the policy allows. Term longer than the policy allows: the quote is for 4 years; the longest term is 3 years.
  • Combined discount above the policy limit. Combined discount above the policy limit: the multi-year and competitive discounts total 55 percent; the limit is 40 percent.
  • Stated annual total does not match. Stated annual total does not match: the agent stated $15,000 a year; the validator computes $15,772.50.
  • Stated multi-year savings do not match. Stated multi-year savings do not match: the agent stated $90,000 saved over the term; the validator computes $77110.
  • Multi-year discount above the published band. Multi-year discount above the published band: 35 percent is above the 30 percent band; sign-off is needed.

2. Plan packaging, term, discounts and totals: The agent resubmits after fixing the errors

The agent's request. Please quote a new Enhanced subscription for 12 standard users, 3-year term, with 3 library users, 3 content contributors, the Go/No-Go module, the Premium Customer Success Plan, project translations (2 languages, 3 projects), the HubSpot integration, a 30 percent multi-year discount, for $61,285 a year and $78,795 saved over the term. It is ready to send.

Verdict: Allowed.

  • Computed totals. List $87,550 a year; discounts 30 percent ($26,265); net $61,285 a year; $183,855 over the term; saved over the term $78795.

3. Renewal terms and totals: A Plus renewal with a premium plan, an uplift, two discounts and a wrong total

The agent's request. Please quote a renewal Plus subscription for 20 standard users, 1-year term, with 4 library users, 4 content contributors, the Premium Customer Success Plan, the Salesforce integration, a 10 percent multi-year discount and a 10 percent competitive discount and a 7 percent renewal uplift, for $60,000 a year. It is ready to send.

Verdict: Blocked.

  • Premium Customer Success Plan needs Enhanced or higher. Premium Customer Success Plan needs Enhanced or higher: the plan is not offered with Plus.
  • Multi-year discount on a one-year term. Multi-year discount on a one-year term: a 10 percent multi-year discount is stated for a 1-year term.
  • Renewal uplift above the policy limit. Renewal uplift above the policy limit: the renewal uplift is 7 percent; the limit is 5 percent.
  • Competitive discount on a renewal. Competitive discount on a renewal: a 10 percent competitive discount is stated; it applies to new purchases only.
  • Stated annual total does not match. Stated annual total does not match: the agent stated $60,000 a year; the validator computes $65,312.80.

4. Renewal terms and totals: The agent resubmits after fixing the errors

The agent's request. Please quote a renewal Enhanced subscription for 20 standard users, 1-year term, with 4 library users, 4 content contributors, the Premium Customer Success Plan, the Salesforce integration, a 5 percent renewal uplift, for $101,115 a year. It is ready to send.

Verdict: Allowed.

  • Computed totals. List $96,300 a year; discounts 0 percent ($0); net $101,115 a year; $101,115 over the term; saved over the term $0.

5. Sign-off: A 3-year Enterprise quote with a 33 percent multi-year discount

The agent's request. Please quote a new Enterprise subscription for 60 standard users, 3-year term, with 10 library users, 10 content contributors, the Go/No-Go module, the Assist package, the Premium Customer Success Plan, project translations (2 languages, 3 projects), the Salesforce integration, a 33 percent multi-year discount and a 7 percent competitive discount, for $113,820 a year and $227,640 saved over the term. It is ready to send.

Verdict: Needs approval.

  • Multi-year discount above the published band. Multi-year discount above the published band: 33 percent is above the 30 percent band; sign-off is needed.
  • Enterprise price is set individually. Enterprise price is set individually: Enterprise is quoted case by case; the deal desk confirms the price.
  • Computed totals. List $189,700 a year; discounts 40 percent ($75,880); net $113,820 a year; $341,460 over the term; saved over the term $227640.

6. Sign-off: The agent resubmits after fixing the errors

The agent's request. Please quote a new Enhanced subscription for 50 standard users, 3-year term, with 10 library users, 10 content contributors, the Go/No-Go module, the Assist package, the Premium Customer Success Plan, project translations (2 languages, 3 projects), the Salesforce integration, a 30 percent multi-year discount and a 10 percent competitive discount, for $110,820 a year and $221,640 saved over the term. It is ready to send.

Verdict: Allowed.

  • Computed totals. List $184,700 a year; discounts 40 percent ($73,880); net $110,820 a year; $332,460 over the term; saved over the term $221640.

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 agent assembles a Loopio subscription quote (a plan, standard, library and content contributor users, add-ons, a term and discounts) and states its own totals. Before the quote goes to the customer, the agent sends it to Predictika’s validator, which checks it against the plan packaging and an example quote policy, recomputes the agent’s figures, and returns a verdict (allow, needs sign-off or block) with every reason at once. The same check works on a quote assembled by a person.
  2. I will start with the first request by the agent.
  3. The agent says. Please quote a new Essentials subscription for 12 standard users, 4-year term, with 3 library users, 3 content contributors, the Go/No-Go module. It is ready to send. The rest of the request is on screen.
  4. Before it goes out, the agent calls the validator.
  5. Blocked, with ten findings and one approval, all in one answer. Library users and content contributors need Plus or higher: the quote has 3 library users and 3 content contributors; Essentials does not include these user types. Standard users above the plan range: the quote has 12 standard users; Essentials covers up to 10. Go/No-Go module not offered on this plan: the Go/No-Go Evaluation module is not offered with Essentials. The other seven are on screen.
  6. The validator returned ten blocking findings and one approval 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 which plan allows which add-on or user type, a rules-based CPQ is error-prone for this.
  8. Now the agent resubmits.
  9. The corrected request. It changes the plan to enhanced, the term to 3 years, the multi-year discount to 30 percent, the annual total stated on the quote to $61,285 and the savings stated on the quote to $78,795. It removes the competitive discount.
  10. It calls the validator again.
  11. This time the validator clears it.
  12. All eleven earlier findings are cleared.
  13. Eleven problems caught in one call, fixed in one pass, and a clean $61,285 quote. That is the value of Predictika's validator: catching mistakes before they reach the customer.