Freddy AI vs Now Assist on Cost

Freddy AI and Now Assist solve similar problems but price them differently, and the difference matters more than the feature comparison when you are forecasting an ITSM AI bill. Freshworks tends toward session and usage-oriented charging for parts of Freddy, while ServiceNow's Now Assist is structured as a per-seat uplift on your fulfiller base. That single structural divergence, usage-leaning versus seat-leaning, changes who overpays, how the bill scales, and which negotiation levers actually work. This comparison is about cost shape, not which assistant is smarter, because two organizations with identical AI needs can end up with very different bills depending on which pricing model they sit under. It is one entry in our complete guide to ITSM AI pricing, which maps all the models side by side.

How each one charges

Now Assist is an uplift: ServiceNow raises the price of your existing fulfiller seats for the AI-enabled tier, so the bill is a function of seat count, fixed regardless of how much the AI is used. Freddy AI, across its copilot and agent capabilities, leans more toward usage and session-based elements alongside tiered packaging, so the bill moves with activity. Neither is inherently cheaper; they fail in opposite directions. A seat-based uplift punishes an oversized base that barely uses the AI. A usage-leaning model punishes heavy, unbounded adoption. Knowing which failure mode you are exposed to is the whole point of comparing them on cost, and on the ServiceNow side that failure mode is driven entirely by seat count, the dynamic our ServiceNow pricing 2026 guide sets out in full.

Who overpays under each model

Under Now Assist, the overpayer is the organization with a large fulfiller base where only a minority will use AI, because the uplift applies to everyone, the dynamic detailed in what the Now Assist uplift actually costs. Under a usage-leaning Freddy model, the overpayer is the organization that drives high assist volumes without a committed-volume discount or a cap, so the meter runs ahead of any budget. The practical implication: a ServiceNow buyer should obsess over seat discipline, while a Freshworks buyer should obsess over usage forecasting and overage protection.

Forecasting the two bills

Because the cost drivers differ, the forecasting method differs. For Now Assist you forecast seats: how many fulfillers will carry the uplift after right-sizing, multiplied by the uplift, across the term. For a usage-leaning model you forecast activity: expected assists or sessions per period, which demands a measured simulation rather than a guess, the discipline in how to run a 90-day AI usage simulation. A buyer comparing the two should run both forecasts on their own numbers, because the cheaper model on paper can be the more expensive one in their specific environment.

The levers differ too

The negotiation moves are not interchangeable. Against Now Assist you shrink the seat base, scope the cohort and cap the renewal uplift, the approach in how to negotiate Now Assist pricing. Against a usage-leaning model you negotiate committed-volume discounts, overage caps and the right to true-down if adoption disappoints. Applying the wrong playbook to the wrong model wastes leverage: haggling a usage rate when the cost is really in your seat count, or shrinking seats when the cost is really in unbounded usage, both leave money on the table.

Choosing on cost, not hype

If you are genuinely choosing between platforms, let the cost shape inform the decision alongside fit and function. An organization with a lean, AI-heavy fulfiller team may do better under a per-seat uplift; one with a large base and selective AI use may do better under a usage model, provided it caps the meter. Either way, the comparison should run on your measured numbers, not the vendors' projections. Running that comparison and negotiating whichever model you land on is the work in our buyer-side AI cost control engagements, fixed fee or gainshare, so we only win when you do.

Frequently asked questions