Pure AI Vendors vs. AI-Enabled Features: Why the Distinction Changes How You Should Negotiate
AI is not one category of spend. It is at least two, with different cost drivers and different levers at renewal.
One distinction I think gets lost in most renewal conversations right now is that "AI" isn't one category of spend, it's at least two very different ones, and treating them the same in a negotiation means missing what's actually driving the cost in each case.
The two categories
Pure AI vendors. Tools whose core product is the AI capability itself, a large language model API, an AI-native writing or coding assistant, a dedicated AI analytics tool. Pricing here is usually tied directly to model usage, tokens, API calls, compute, and the vendor's own primary cost driver is the underlying model infrastructure.
AI-enabled features within existing SaaS tools. An established platform, a CRM, a support tool, a project management tool, that has added AI capabilities into its existing seat-based product, often as a premium add-on or bundled into a higher tier.
One budget line, two different negotiations
- Cost driver
- Tokens, calls, compute
- Moves with
- How the tool is used
- The real lever
- Volume tiers and caps
- Cost driver
- Tier upgrade or add-on
- Moves with
- Seats on the higher tier
- The real lever
- Is the feature used at all
Pushing back on "the AI pricing" without knowing which column you are in aims at the wrong lever in both cases.
Why this distinction matters for negotiation
For a pure AI vendor, the pricing conversation is fundamentally about usage and consumption, tokens, API volume, model tier, and the negotiation should focus there. For an AI-enabled feature bolted onto an existing SaaS tool, the more relevant question is usually whether the feature justifies the tier upgrade or add-on cost at all, separate from the base product's value, since the AI capability there is often a pricing lever more than a core product shift.
The trap: treating both the same way
I've seen renewal conversations where a buyer pushes back generally on "the AI pricing" without distinguishing which category they're actually dealing with. For a pure AI vendor, that vague pushback misses the real lever, the usage-based cost structure. For an AI-enabled feature, it misses the real question: is this feature, specifically, worth the incremental cost, independent of the base product you already value.
What to check before your next renewal
- For each vendor, is the AI component the core product, or an add-on to something you're already paying for
- If it's an add-on, is it being used enough by your team to justify the incremental cost specifically
- If it's the core product, is the pricing structure usage-based, and do you understand what drives that usage