Freddy AI Pricing and What It Adds to Your Bill

Start with the structure, because the pricing follows it. Freshworks splits Freddy into two products. Freddy AI Copilot is the agent-facing assistant that drafts replies, summarises tickets and suggests next steps, priced as a per-seat add-on on top of the agent subscription. Freddy AI Agent is the autonomous, requester-facing layer that resolves issues without an agent, priced on a consumption basis, per session or per resolution. A little lightweight capability is folded into the higher plan tiers, but the features most buyers actually want are paid add-ons. For the full pricing context, see our Freshservice pricing guide for 2026.

The two layers and how each is metered

The two layers fail in different ways. Copilot fails on scope: attach it to every agent and you pay for a per-seat assistant that most agents open once and forget. Freddy AI Agent fails on volume: sign an open-ended session model and the line scales with traffic you do not control. The discipline that protects you is the same one we apply to every AI line across platforms, set out in our guide to ITSM AI pricing.

What Freddy actually adds to the bill

Run the arithmetic on Copilot first. A per-seat add-on attached to a full agent base can rival a tier upgrade in absolute cost, which means it deserves the same scrutiny as the tier decision itself. The honest question is not "is Freddy useful" but "how many of our agents will use it enough to justify the seat." In most estates that number is a fraction of the base, concentrated in the high-volume queues, which is exactly the scope to license.

Freddy AI Agent is the harder line to forecast, because its value and its cost both depend on deflection. If it resolves a meaningful share of requester traffic without an agent, it can pay for itself; if it resolves little, you are paying a consumption meter for a feature in search of a use case. Either way the contract needs a cap and a clear unit price so the meter cannot run away between renewals. Treating the AI line as a separate negotiation with its own term and exit, rather than a rounding error inside the subscription, is the single most important move.

How to scope Freddy before you sign

Pull Copilot usage by agent if you already have it, or pilot it with a defined group if you do not, and license the seats that the data supports rather than the whole base. For Freddy AI Agent, model realistic session volume against your actual requester traffic, not the vendor's demo deflection rate, and negotiate a cap with a unit price you understand. Keep both layers on the shortest term Freshworks will allow, so you can re-price once you have real usage rather than locking a multi-year rate against a guess.

This scoping work is part of a broader renewal sequence. The full playbook is in how to negotiate a Freshservice renewal, and when AI is the dominant line, our AI cost control service runs the modelling and the negotiation. The commercial hub for the platform is the Freshservice platform page.

The questions to ask Freshworks before you commit

The AI line is easiest to negotiate while it is still optional, which means the questions matter most before anything is signed. Ask Freshworks to state, in writing, exactly how each layer is metered: is Copilot billed per active seat or per provisioned seat, and does Freddy AI Agent charge per session opened or per issue resolved. The difference is not academic. A per-provisioned-seat Copilot charges for agents who never open it, and a per-session Freddy AI Agent charges for conversations that never deflect. Pin the unit down and you can model the cost; leave it vague and you are signing a blank meter.

Then ask what happens at the edges. What is the overage rate once a consumption allowance is exhausted, and is it the same as the committed rate or a punitive one. Can unused Copilot seats be reallocated mid-term, or are they locked for the year. Does the AI pricing hold at renewal, or is it exposed to the same annual uplift as the subscription. Each answer is a term to negotiate, and the right to re-scope AI at renewal is one of the most valuable, because usage data from year one almost always argues for a different shape than the one you bought on day one.

Finally, weigh Freddy against the alternative of not buying it yet. AI that is genuinely deflecting tickets or accelerating agents earns its place quickly; AI bought on the promise of future productivity is a cost looking for a justification. There is no penalty for scoping small, proving the value, and expanding from evidence, and there is a real penalty for the reverse. This is the same posture we take on every platform, and it is why the renewal sequence in how to negotiate a Freshservice renewal treats AI as its own line rather than a rounding error inside the subscription.

Frequently asked questions

How is Freddy AI priced?

In two layers: Freddy AI Copilot as a per-seat add-on for agents, and Freddy AI Agent on a per-session or per-resolution consumption basis for autonomous resolution. Some light capability is bundled into the higher tiers, but the main features are paid add-ons.

What does it add to a Freshservice bill?

Copilot adds a per-seat charge that, across a full base, can rival a tier upgrade. Freddy AI Agent adds a consumption line that scales with volume. The overspend is attaching Copilot to everyone and signing an uncapped session model.

How do I control Freddy AI cost?

Scope Copilot to real users, confirm the Freddy AI Agent metering, model realistic volume, cap consumption with a clear unit price, and keep the AI line on a short term so it can be re-priced once you have usage data.