---
name: wavicle-ai-project-scoping
description: Use when a user wants to evaluate whether Wavicle's published AI consulting services fit a proposed project using public service, pricing, readiness, and case-study sources, without submitting forms or claiming a booking.
---

# Wavicle AI project scoping

## When to use this skill

Use this skill when a user wants to:

- compare a proposed AI product, automation, integration, or deployment with Wavicle's published services;
- prepare questions for a Wavicle discovery call;
- identify a relevant Wavicle-published case study;
- build an initial, non-binding scope, delivery-range, or readiness checklist from Wavicle's public material.

Use the skill when the task needs judgment across several published sources,
especially when service fit, evidence, delivery constraints, and next questions
must be assembled into one brief. For simple structured lookup, prefer the
[read-only public API](https://www.wavicle.tech/api/v1) or the
[read-only MCP server](https://www.wavicle.tech/.well-known/mcp/server-card.json).
Do not use the skill for unrelated vendors or as a substitute for an executed
statement of work.

Do not use this skill to:

- claim that Wavicle accepted a project, issued a quote, sent a message, or booked a call;
- treat a Wavicle case-study outcome as independently verified;
- promise a price, timeline, technology, integration, compliance outcome, or service that is not stated on a cited page;
- imply that the public API or MCP server can write data, send an inquiry, create a quote, book a call, authenticate a client, or operate as a sandbox.

## Authoritative public sources

Start with the smallest relevant set:

1. [Service overview](https://www.wavicle.tech/services.md)
2. [Pricing](https://www.wavicle.tech/pricing.md)
3. [Case-study index](https://www.wavicle.tech/case-studies.md)
4. [AI readiness assessment](https://www.wavicle.tech/resources/tools/ai-readiness.md)
5. [AI project cost calculator](https://www.wavicle.tech/resources/tools/ai-cost-calculator.md)
6. [About Wavicle](https://www.wavicle.tech/about.md)
7. [Contact Wavicle](https://www.wavicle.tech/contact.md)
8. [Public resource API](https://www.wavicle.tech/api/v1)
9. [OpenAPI 3.1 document](https://www.wavicle.tech/openapi.json)
10. [MCP server card](https://www.wavicle.tech/.well-known/mcp/server-card.json)

Use [llms.txt](https://www.wavicle.tech/llms.txt) as the concise index. Fetch
[llms-full.txt](https://www.wavicle.tech/llms-full.txt) only when the linked
pages do not provide enough context.

## Workflow

1. Restate the user's desired business outcome in one sentence.
2. Record the current workflow, volume, owner, inputs, outputs, exceptions,
   sensitive data, systems involved, and measurable baseline.
3. Match the request to a service only when the published service page supports
   the match. Otherwise label the fit as unknown.
4. Select at most three relevant case studies. Cite each canonical HTML URL and
   label its numbers as Wavicle-published outcomes.
5. Use the published pricing page or calculator only for a non-binding range.
   State which inputs or assumptions drive that range.
6. Identify missing discovery inputs, delivery risks, human-review gates, and
   acceptance criteria.
7. Present a short brief for the user to review before they contact Wavicle.

## Output

Return:

- desired outcome and measurable baseline;
- likely service fit, with citations;
- relevant published evidence, clearly attributed;
- known inputs, systems, data, and constraints;
- open questions and risks;
- non-binding price or timeline context only when sourced;
- suggested next step requiring explicit user action.

## Human action boundary

The public site offers email and an embedded Cal.com scheduling experience.
Those are human-reviewed actions. Give the user the canonical
[contact page](https://www.wavicle.tech/contact), but never report that an
inquiry or booking occurred unless the user completes it and receives
confirmation.