A company needs its people to work together. We want the decision to include another person to follow the work they do, rather than the price of another software seat. That is the starting point for the platform model we are building toward.
The planned model
Our intended platform base is $200 per company per month, with unlimited users and $50 in monthly credits shared across hosting and intelligence usage. Additional credits are purchased when the included allowance is used.
This is a planned model for the platform, still in development. The base fee gives the company a common place to work; the allowance contributes toward the variable resources it consumes. Unlimited users means no per-seat charge, while hosting and intelligence remain usage-based.
Keep participation separate from consumption
A person who checks a record occasionally should be able to participate without becoming another license calculation. The same applies to a manager reviewing information or someone helping correct an entry. Their contribution may be important even if they spend little time inside the product.
Processing a large collection of documents is a different kind of event. It uses storage, compute, and potentially model inference. A company that chooses more of that work consumes more resources. Keeping these two ideas separate makes the commercial direction easier to explain: the team can participate, and usage has an allowance and a cost.
The allowance is shared
The $50 allowance is one monthly credit pool for hosting and intelligence together. It is not $50 of hosting plus a separate $50 of model usage. Its purpose is to include an initial amount of running and processing work within the company base.
A usage-based model also needs an understandable account of consumption. Our platform work includes company-level billing, metering, and configurable controls for intelligence spending. The product needs to explain the balance, the work drawing against it, and what happens when a limit or balance is reached.
The exact amount of processing a credit buys depends on the task and resources involved. We should explain those units in the product where someone decides to use them. A flat company base should not become a vague promise of unlimited AI.
Intelligence should justify the activity
We do not want intelligence to be a feature label attached to every screen. It needs a useful job: finding the passage behind a question, preparing something a person can review, or helping move a workflow forward with the right context.
Charging for usage makes that usefulness especially important. The person requesting work should have a reason to believe the result will help. The system should make it possible to inspect the result and understand its limits. An impressive answer that cannot be checked is a weak foundation for an operational decision.
The application still needs to be dependable
Accessible pricing is a commercial choice, not an excuse to treat the application as disposable. Accounts, permission checks, document history, forms, and ongoing support remain real engineering work. Our shared platform and design system are ways to carry that work forward across products.
The bet is that we can make the conventional application layer increasingly affordable while building a business around useful intelligence. We will need to prove that balance through actual product use and operating costs.
Start with a defined industry and useful stages
Construction is our first industry focus. Inventory management and time tracking come first; presales, including estimating and takeoffs, follows. Each stage gives us a concrete set of workflows in which to test what the foundation provides and where intelligence earns its place.
The commercial direction belongs alongside that product discipline. People should understand what access covers, which resources are metered, and what useful work they are choosing to pay for. That is the model we want to make real.