---
title: "What you can use Pave Agent for"
description: "What you can use Pave Agent for"
canonical_url: "https://support.pave.com/articles/what-you-can-use-pave-agent-for-kbywRhU3qY"
md_url: "https://support.pave.com/articles/what-you-can-use-pave-agent-for-kbywRhU3qY.md"
---
# What you can use Pave Agent for

## What is this article?
Pave Agent answers compensation questions in conversation, using your company's data in Pave and Pave market data. If you have access and are not sure where to start, this article gives you a set of questions worth trying.

This is a starting set, not a complete list of what Pave Agent can do. Questions outside these patterns are worth trying too. Use the examples as a starting point and change the roles, levels, and locations to match your organization.

[Screenshot: The Pave Agent chat with an example question typed into the message box]

## Offers and employee comp analysis
Questions about a single person or a single decision: what to offer, what to pay after a move, or whether someone's pay holds up.

**Evaluate an offer.** Give it the role, level, location, and the number you are considering.
> "Looking to make an offer to a Software Engineer P4 in London of $180k base, please evaluate."

Follow up with the rest of the package:
> "Recruiter wants to offer $72k equity vesting over 3 years with no cliff."

**Evaluate an offer against a candidate's background.** Attach the resume to your message instead of typing out the details.
> "Recruiter wants to offer this candidate $250K base, please evaluate this offer."

It reads the role and location from the resume, so you do not have to specify them, and uses the candidate's background to check the leveling.

**Price a relocation.**
> "[Employee name] is moving to London, no change in role. What should their new compensation be?"

Follow up on the parts that relocation policies usually leave open:
> "How should we handle their equity in this move? Does relocation change anything, and how does their total annual comp compare to London market afterward?"

**Investigate a comp complaint.**
> "[Employee name]'s manager thinks they're underpaid, can you investigate? Include their performance, raise, and promotion history."

## Market pricing and pay bands
Questions about how your pay compares to market, across a team, a slice of the organization, or a role you have never hired before.

**Benchmark an entire org.**
> "Benchmark the entire Data Science org rolling up to [leader name]. Would like it to cover all pay components and all locations."

Then price the gap:
> "How much would it cost to bring all these employees to at least P50 base?"

**Check competitive positioning on a specific slice.**
> "What is our market positioning on new hire equity for Software Engineers in San Francisco?"

It pulls the relevant employees and ranges, pulls market data (Pave market data or your own benchmarks in Market Pricing), matches your internal data to market on role, level, and pay zone, and returns a prioritized snapshot. This works best once your peer group and target percentile are in your Guidance.

**Compare peer groups against each other.**
> "Can you compare market data benchmarks and data quality across my 5 peer groups? Focus on P75, San Francisco, Software Engineering P levels."

**Benchmark a niche role using job postings.** Useful for roles that do not map cleanly to a survey job family or title.
> "I'm hiring a Propulsion Engineer in LA, use both Pave market data benchmarks and job postings to inform how much I should pay for this role."

> "Show me those postings."

It finds semantically similar titles in the job postings data and proposes the closest matches in the Pave taxonomy, so you can compare a niche role against official benchmarks.

**See what peers are hiring for.**
> "What roles is [company name] hiring for right now?"

## Merit and comp cycle planning
Questions that use your merit cycles in Compensation Planning, either an active cycle or a finished one.

**Model merit budget scenarios.**
> "We're setting up our Q3 2026 cycle, can you model out what our rec logic should be if we have 3% vs 4% budget? Use the Q3 active cycle for context on eligible employees and current cash spend, and assume performance rating distribution and promotion rate looks similar to Q1 2026 cycle."

Point it at a live cycle for eligible employees and current spend, and at a past cycle for rating distribution and promotion rate.

**Recap a finished cycle.**
> "Summarize my last finalized cycle in terms of performance rating distribution and raises."

**Allocate discretionary budget.**
> "Please recommend individuals we should consider spending discretionary cash or equity on. For each person, give me your rationale and the amount you'd recommend."

Most of a cycle is rules-driven. Discretionary budget is not, and it is usually allocated last with the least analysis behind it. It reasons across compa ratio, time since last adjustment, tenure and performance inputs where available, and equity vesting position. Treat the result as a first-pass shortlist.

## Comp structure and modeling
Questions about the shape of your pay structure and what it would cost to change it.

**Cost to minimum and range penetration.**
> "How are my employees distributed within base salary ranges?"

> "What is the cost to bring all US employees to range minimum? Now show cost to 15% range penetration, detailed by function."

**Model a job architecture change.**
> "Analyze the impact of consolidating our brand marketing and product marketing job families into a single band, that lands halfway between the existing bands. Which employees would fall out of the new band?"

**Benchmark span of control.**
> "Show me number of direct reports per Director. Then aggregate by department and compare against market median span of control."

This is an analysis over your employee data, which includes reporting relationships, compared against market benchmarks on span of control. Use it to find where you are top-heavy or where managers are overloaded relative to market.

**Choose a hiring location.**
> "I'm hiring a Machine Learning engineer for the first time, where should I consider, taking into account my existing employee locations, MLE talent density, and geo-differentials?"

It weighs your existing footprint, your geo pay policy, market geo differentials, and how employees in Pave's market data are distributed across locations.

## Equity
**Benchmark equity participation.**
> "What percentage of P1 employees in the US in Pave's dataset receive new hire equity?"

> "How does our equity participation compare to market?"

There is a second version of this question that sounds similar and answers something different:
> "What percentage of companies in Pave's dataset grant new hire equity to [level or function]?"

The first asks what share of employees receive equity. The second asks what share of companies grant it. Pick the one you mean.

## Research, reporting, and setup
**Search Pave Data Lab.**
> "Search Pave Data Lab for conversations about how I should think about off-cycle promo equity grants."

Pave Data Lab includes Pave's own analysis alongside pulse surveys, polls, and community discussions from compensation leaders. Asking it to search there tells you how other practitioners approach a question. This searches published research rather than calculating a figure from data.

**Export a custom data report.**
> "Create a CSV export of engineering ranges including job code, pay zone, range midpoint, market median for base salary, and average compa ratio."

Describe the population, the columns, and the grouping in plain language. See the article on uploading files and exporting results for how downloads work.

**Draft a new Skill.**
> "Help me write a new skill for identifying bands out of sync with market."

It interviews you about the use case and drafts the Skill content. It can also walk through a dry run as a final check if you ask for one. The draft is copy and paste into a new Skill. See the article on setting up Skills.

## Things to know
**Wide-scope questions take longer.** Benchmarking an entire org across every pay component and location takes a few minutes. If you do not need the full picture, narrowing to one function, one region, or one level range keeps the request smaller.

**Questions about your own organization work from what is loaded in Pave.** If your data is not in Pave, Pave Agent cannot analyze it. Pave market data, job postings, and Pave Data Lab research do not depend on your own data being loaded. See the article on what data Pave Agent can see.

**Answers improve once your Guidance is set up.** Your positioning metric and your benchmarking approach stop needing to be restated in every question. See the article on setting up Guidance.

**Some of these workflows have a matching built-in Skill.** Pave Agent selects the relevant one based on what you ask, so there is nothing to choose. See the article on setting up Skills.