I have sat in the room when AI gets sold this way, and the trick is the word "included." Something is always included at a base level, and the upgrade that actually does the work is the one that meters. Copilot and Agent are not the same purchase, and pretending they are hands the vendor a quiet expansion at every renewal. Separate them and the negotiation gets simple.
Start here: Book an AI cost control review
What is the difference between Freddy Copilot and Freddy Agent?
Freddy Copilot sits next to a human. It drafts replies, summarizes long tickets, suggests next steps and speeds up the agent who is still doing the work. Because it attaches to a person, it is typically priced per agent seat. Freddy Agent removes the human from the simple cases. It handles a request end to end, autonomously, so it is typically priced on usage or sessions rather than on seats. One scales with headcount, the other scales with volume. That single distinction drives everything about how each one hits your bill.
| Dimension | Freddy Copilot | Freddy Agent |
|---|---|---|
| What it does | Assists a human agent in the console | Resolves requests autonomously |
| Typical pricing unit | Per agent seat | Usage or sessions |
| How the bill grows | With the number of agents licensed | With the volume it deflects |
| Forecastability | Predictable, scales with headcount | Variable, scales with adoption |
| Main buyer risk | Licensed to agents who never use it | Uncapped consumption at renewal |
As of 2026, the exact unit pricing varies by Freshservice plan tier and deal size, so confirm both numbers against your order form rather than the vendor's summary slide. The published plan tiers as of 2026 run Starter $19, Growth $49, Pro $95 and Enterprise $119 per agent per month, and Freddy scope shifts across them, which is another reason to pin the AI line separately.
Which one actually inflates the bill?
Both, in different ways, and the buyer mistake is to worry about the wrong one. Copilot inflates the bill when it is licensed to the whole team and used by a fraction of it, the same waste as any per-seat over-provision. Agent inflates the bill when deflection climbs and the usage meter has no ceiling, so the better it works, the more it costs. Copilot is a right-sizing problem. Agent is a capping problem. You solve them with different moves.
How do I keep Freddy from inflating my renewal?
Three moves, run before the quote is final. First, scope Copilot to the agents who actually invoke it, not the full roster, the same seat discipline covered in how to negotiate a Freshservice renewal. Second, put Agent on a short term so you can re-price it once you have real usage data, because year-one optimism is not a budget. Third, negotiate the consumption cap and the overage rate up front, so a busy quarter cannot turn into a surprise top-up.
The same pattern repeats across the agentic AI category. An AI-first platform meters differently again, as covered in Atomicwork pricing and the AI-first ITSM cost model, and the cross-vendor playbook sits in the 2026 guide to agentic AI ITSM pricing and negotiation, which is the hub for this cluster.
Our gated Freshservice Buyer Guide includes the Freddy scoping worksheet that splits Copilot seats from Agent usage so each is negotiated on its own terms.
What to put in the contract
Write the AI scope as its own section. Name the Copilot seat count and tie it to a usage review so you can drop unused seats at renewal. Define the Agent unit, whether it is a session or a resolution, in plain words with edge cases listed. Set a not-to-exceed ceiling on Agent consumption and a fixed overage rate. And add a re-bundling protection so Freshworks cannot quietly re-price the AI scope at the next renewal. This is the buyer-side discipline the firm runs across more than 500 engagements, $420M of negotiated contract value and a 30% average reduction. The Freshservice commercial hub is the Freshservice platform page, and the engagement runs through AI cost control on fixed fee or gainshare.
Frequently asked questions
- What is the difference between Freddy Copilot and Freddy Agent?
- Copilot assists a human agent in the console and is typically priced per seat. Agent resolves requests autonomously and is typically priced on usage or sessions. They add to a Freshservice bill differently, so scope and negotiate them separately.
- Which Freddy tier costs more?
- It depends on volume. Copilot scales with how many agents you license. Agent scales with how many requests it handles autonomously, so a high-deflection rollout can cost more than expected. Model both against your real ticket data.
- How do I keep Freddy from inflating my renewal?
- Scope Copilot to the agents who use it, put Agent on a short term so you can re-price it, and negotiate the consumption cap and overage rate up front. Confirm 2026 unit pricing against your order form.
Split the Freddy line.
We separate Copilot seats from Agent usage, right-size one and cap the other, before you sign. Fixed fee or gainshare.
Book an AI cost control review →