# Terrabase

> Terrabase is the decision runtime for recurring operations. It runs decisions like pricing, ordering and next best action end to end: a signal opens the decision, your policies govern it, your people approve or change it, the action executes, the outcome is measured, and every result makes the next decision sharper.

Most operational decisions repeat. A rep decides what to offer on every call. A store manager decides what to order every week. A pricing team decides what to charge every month. Today those decisions live in dashboards, spreadsheets and individual judgment, and the reasoning disappears the moment they are made. Terrabase keeps the whole loop: what fired, what was decided, why, what it cost, and what it returned.

## How it works

Every decision runs through the same loop:

1. Signal. Something changes: a competitor moves price, rain is forecast, a customer's contract nears its end. The signal opens a decision.
2. Decision. Terrabase ranks the options under your policies and picks one, with the reasons written down.
3. Action. The chosen move goes out to the system where work happens. Your operator approves it first, and can change it before approving.
4. Outcome. The result is measured against what was expected, in dollars and volume, not vibes.
5. Learning. What worked becomes tomorrow's default. What an operator changed becomes a rule the system remembers.

Operators stay in charge. When an operator changes a suggestion before approving it, Terrabase records the change and learns the pattern, so the next suggestion starts from what the operator knows.

## What it runs (example scenarios shown on the site)

- Next best action for a telecom call center: one day on the floor. Reps get a ranked, policy-checked offer for each customer before the call, pitch it, and the runtime records which offers landed. One rep changes a suggested plan to a shorter term with a smaller data pack; the customer says yes, and the system learns that pattern for similar customers.
- Daily operating plan for a quick service restaurant: one week around a soccer final. Monday plans the week including the match and approved staff leave. A rain forecast on Wednesday moves the plan. Saturday's foot traffic triggers last-hour changes. Sunday saves the plans that worked.
- Monthly price for a retail and consumer goods brand: ten days of a pricing cycle. A competitor raises price 6 percent. Terrabase finds the smallest raise worth making, 4 percent, which keeps us cheaper, holds volume and adds 17 percent gross margin. It answers a competitor promotion only on the channel where it runs, prices a marketing campaign region by region, and rolls back three cities where the raise did not hold.

## Investigate any decision

Every decision on the page can be questioned: why was this option chosen, which rule fired, what was excluded and why. Answers come with evidence. The evidence can be exported for an audit, and any run can be replayed end to end.

## Other products

- Terrabase Analytics (https://terrabase.ai/analytics): an analytic operating layer. A self-learning context layer maps and maintains your analytics knowledge, and an analytics harness turns plain-language questions into reports, dashboards and answers you can trace to source.

## Contact

- Email: support@terrabase.ai
- Request a demo by leaving your email on https://terrabase.ai/