Vai by Variabl

AI that knows how your comp actually ran

Vai is the AI analyst built into Variabl’s calculation engine. Comp admins use it to trace variances and check changes before they go in. Sellers use it to get answers instead of filing tickets. With the Variabl MCP server, your team can ask Vai from Claude, ChatGPT, or your own internal agent.

Get a demo Connect Vai to your AI

“What used to take days now takes less than ten minutes.” Sarah Frongillo, NetNumber.

Last updated November 5, 2026

Vai investigation findings explaining why AE FY26 paid $74,727 more in February than in January, broken down by component

What is Vai?

Vai is Variabl’s AI analyst for sales compensation. It lives inside the calculation engine, so it knows how the system works, plus your plans, your data, and the history of every calculation run: what ran, when, for whom, and what changed. Admins and sellers ask Vai questions in plain English and get answers backed by the records and rules behind them.

How is Vai different from AI in other commission software?

Vai works from the same plans, transactions, and run history the calculation engine uses, so it can explain how a specific payout was calculated. Most AI in commission software is bolted on. It’s a support bot trained on the user manual and past tickets, or a copilot reading report outputs. It can tell you how the system is supposed to work. It can’t tell you how your comp actually ran.

Variabl stores compensation logic as structured data from the ground up. That’s why Vai can explain a number, not just describe a feature.

Question Bolt-on AI Vai, built in
What it reads User manual, past tickets, report outputs Plans, transactions, configuration, and every calculation run
What it can answer How the system is supposed to work Why a specific payout came out the way it did
What it can do next Point to a help article Recommend the setup fix, build the report, or flag a process gap

What can comp admins do with Vai?

Comp admins use Vai to trace variances, check changes before they go in, and get answers, reports, and fixes in plain English. Vai turns the busiest week of the month into a conversation with your data.

Trace a variance in minutes

Ask Vai why an accrual differs from what’s payable. It traces the variance run by run, by date and plan component, and explains it. It also flags what nobody asked about, like a test run that landed in a real pay period, and recommends the setup fix.

Vai in Variabl explaining why the AE_FY26 plan paid $74,727 more in February than January, broken down by plan component
Vai traces a $74,727 month-over-month change to the plan component that drove it.

Check a change before you make it

One customer had a full quarter of transactions loaded into a single month. They needed to spread them across three months by actual close date, so sales leaders got the right indirect credit as territories shifted. Many of those transactions already carried manual overrides. Before touching anything, the team asked Vai whether the overrides would survive the move and what the safest path was. Vai worked out how the transactions and overrides were stored and how attainment and rates were calculated, then laid out a way to restructure the quarter without losing any of the customer’s work.

Get the answer, the report, or the fix

Day to day, admins ask Vai questions in plain English and get an answer, a new report, or a recommended process fix.

Vai in Variabl summarizing total paid per deal with a breakdown by person, delivered as downloadable files
Ask for a report in plain English and Vai builds it, with the files attached.

How does Vai help sellers understand their pay?

Sellers ask Vai about their commissions, earnings, payments, or compensation in plain language and get the answer in the moment, instead of waiting on comp ops. Sellers who understand their pay act on it.

  • Answers, not tickets. Real disputes become a case inside Variabl, with an owner and a status until it’s resolved.
  • Every number has a trail. Statements break out credits, commissions, bonuses, draws, and balances by period, and each figure traces back to the deals and rules behind it.
  • On web and mobile. Sellers reach their dashboard, statements, cases, and Vai on the web or the Variabl app for iOS and Android, with SSO.

Fewer questions reach your comp team, and the field trusts the number.

What is the Variabl MCP server?

The Variabl MCP server connects Vai to your company’s AI tools through the Model Context Protocol, an open standard for connecting AI applications to outside systems. Your team can connect Variabl to Claude, ChatGPT, or an internal agent, then ask comp questions where they already work.

Most MCP connections hand raw data to your model and leave it to guess what the numbers mean. Ours runs the investigation inside Variabl first, with the full comp context, and checks the supporting data before it answers. Your AI gets the explanation and the supporting files, not a data dump.

How the connection works

  1. 01

    Approve

    A Variabl owner or administrator approves the MCP connection.

  2. 02

    Ask

    Your AI tool sends a question to Vai.

  3. 03

    Investigate

    Vai investigates inside Variabl, across the relevant plans, configuration, records, and results. Investigations typically take several minutes.

  4. 04

    Answer

    Vai returns its explanation and supporting files, and your AI can combine them with other sources your company has connected.

What teams do with Vai over MCP

  • Explain a disputed payout with a trail back to the source records and plan logic.
  • Reconcile commission results against CRM, crediting, and quota data.
  • Combine comp with pipeline, revenue, budget, and activity data, such as commission expense as a percentage of revenue.

Is it safe to connect commission data to ChatGPT or Claude?

With the Variabl MCP server, your administrators decide who can connect, access runs through SSO, and your AI sees only the sources your company has authorized. Other systems, such as a CRM or finance system, connect separately under their own permissions.

  • A Variabl owner or administrator approves every MCP connection.
  • What each person sees follows their permissions in Variabl.
  • AI explains and recommends. The deterministic calculation engine does the arithmetic, and people approve every plan change.

Questions you can ask Vai

  • Why did my commission change from last period?
  • How do the tiers, eligibility, and payment rules in this plan work?
  • Is this commission hold consistent with our invoice-paid policy?
  • Which fields caused these transactions to be excluded?
  • Why does accrued commission differ from what’s payable this month?
  • What was commission expense as a percentage of revenue last quarter?

Where does Vai work by industry?

Vai works across every industry Variabl serves, because it reads each customer’s own plans and data. In insurance, for example, Vai reconciles carrier commission statements, forecasts contingent commissions, and explains producer disputes. Read more about insurance commissions.

Watch Vai work through your scenario

The fastest way to judge Vai is to see it answer a question your team deals with today. Bring a real plan and a real problem, and we’ll walk through it together.

This field is for validation purposes and should be left unchanged.

Variabl intro and product overview

Zoom

_ga
gclid
msclkid
li_fat_id

Vai and Variabl MCP FAQ

Vai is the AI analyst built into Variabl, an AI-native incentive compensation management platform. It answers sales compensation questions for admins and sellers using the plans, data, and calculation history inside Variabl.

The most useful AI for sales compensation works inside the calculation engine, so it can explain how a specific payout was calculated rather than describe how the software should work. Vai, Variabl’s AI analyst, is built that way. It traces variances run by run, answers seller questions in plain English, and connects to Claude, ChatGPT, or internal agents through the Variabl MCP server.

Vai works inside Variabl’s calculation engine and can see the history of every calculation run. Bolt-on assistants are trained on documentation or read report outputs, so they can describe how a system should work but can’t explain how a specific payout was calculated.

Yes. Sellers ask Vai about their commissions, earnings, payments, and compensation in plain language and get answers without contacting an administrator. What each seller sees follows their permissions in Variabl.

An MCP server lets an AI application, such as Claude or ChatGPT, connect to an outside system through the Model Context Protocol, an open standard. The Variabl MCP server is the MCP server for Variabl’s incentive compensation platform, and it lets an AI tool request investigations from Vai.

Yes. The Variabl MCP server connects Vai, Variabl’s AI analyst, to Claude, ChatGPT, and other MCP-compatible platforms. Vai runs each investigation inside Variabl with full compensation context and returns an explanation and supporting files.

Claude, ChatGPT, and other MCP-compatible platforms, including internal agents a company builds. Check with Variabl to confirm compatibility for a specific tool.

No. A Variabl owner or administrator approves the connection, access runs through SSO, and your AI sees only the sources your company has authorized and connected.

No. Variabl’s deterministic calculation engine does the arithmetic. Vai explains results, traces variances, recommends setup fixes, and builds reports. People approve every plan change.

Most Vai investigations take several minutes. Some may need more information or help from Variabl support.

The Variabl MCP server is available to participating organizations. Contact Variabl to discuss access.

Yes. In insurance, Vai reconciles carrier commission statements, forecasts contingent commissions, and explains producer statements, using the same engine Variabl uses for carriers, agencies, brokerages, MGAs, and IMOs.