Am I the only one who thinks their pricing model punishes active researchers?

I've been meticulously reviewing SciSpace's pricing page and usage logs from my team's account over the last quarter, and a concerning pattern has emerged. Our monthly charges have become increasingly volatile and unpredictable, not because our research output has changed dramatically, but due to the sheer volume of our iterative querying and literature review processes. The core issue appears to be the shift from a model based primarily on subscription tiers to one heavily weighted by consumption of "AI Words" or similar compute units.

From an audit-logging perspective, this creates a problematic scenario where cost becomes directly tied to investigative diligence. Let me illustrate with a concrete workflow that now carries a direct financial penalty:

* A researcher formulates a complex query to find papers on a niche topic.
* They receive a response, then naturally ask five follow-up questions to clarify, dig deeper into methodologies, or request summaries of specific sections from the returned papers.
* This iterative dialogue, which is the essence of thorough research, is now a metered event. Each refinement, each request for "more detail on figure 3," consumes from the same prepaid pool.
* Compare this to a more passive user who runs one broad query per paper. Their costs are minimal, but their engagement and derived value are also surface-level.

This feels antithetical to the goal of a research assistant. It's as if we're being billed per "thought" or per "question" in a literature review session. I can trace this directly in our logs. A single, deep-dive research session can generate hundreds of individual log events (queries, follow-ups, copilot actions), each translating to a micro-debit from our balance.

Furthermore, from a compliance and budgeting standpoint (SOX, anyone?), this is a nightmare. Forecasting monthly costs for an active research team is nearly impossible. One month, a team might be in an exploratory phase, generating high query volumes. The next, they might be in a writing phase, with low consumption. The bill fluctuates wildly, not based on the *value* of output, but on the *process* of input.

I have to ask: does this model not inherently discourage the very activity SciSpace is built to support? Deep, iterative, question-driven exploration is the cornerstone of research. Putting a direct per-unit cost on each step of that conversation seems to punish the most engaged and curious users. I'm curious if others have analyzed their usage logs and observed similar disconnects between user activity spikes and perceived value versus the incurred costs. Are teams now having to implement internal "query budgets" or discourage follow-up questions to stay within forecast?