How Freelancers Can Use Claude Skills to Turn a Discovery Call Into a Statement of Work

If Claude Skills is new to you

Claude Skills are reusable instruction folders that Claude can load when a task matches the skill description. Current Anthropic help says Skills are available on Free, Pro, Max, Team, and Enterprise plans when code execution is enabled.

For this workflow, you will create one private custom skill, upload it in Claude under Customize > Skills, enable it, test that Claude recognizes it, then use it to turn discovery-call material into a first-draft statement of work. The Claude Skills tool hub is useful further reading, but this page gives you enough setup detail to run the workflow.

This is not legal advice and not a substitute for contract review. Claude should organize the SOW draft, preserve open questions, and keep scope boundaries visible. It should not invent pricing, milestones, timelines, legal terms, or included work.

What skill you will use in this workflow

You will use a custom skill called sow-from-discovery-call.

That skill does one narrow job: it turns a discovery call transcript, intake form, and onboarding notes into a first-draft statement of work. It extracts confirmed scope, deliverables, assumptions, exclusions, milestones, open questions, and approval terms into a reusable SOW shape.

This task is repetitive enough to deserve a skill because freelance SOWs often use the same sections even when the client and project change. Before using the skill, prepare the transcript or summary, intake notes, client context, engagement rules, and any pricing or approval terms you already confirmed.

Problem and who this is for

A discovery call goes well, then the real work starts. You need a statement of work that reflects what the client actually asked for, what you actually agreed to, and what still needs definition. If you rush, you either overscope the job or send something vague that creates revision pain later.

This workflow is for freelancers, agencies, consultants, and solo operators who run discovery calls and need a usable statement of work fast.

Prerequisites

• Claude with code execution enabled and access to Skills.
• A call transcript or a solid call summary.
• Your preferred SOW sections. Example: scope, deliverables, timeline, assumptions, exclusions, revision policy, investment, approvals.
• A decision on what you will not promise without follow-up.
• Permission to use the client material in Claude under your contract terms.
• A human review step for pricing, legal language, scope boundaries, and approval terms before anything is sent.

How to gather the source material

1. Get the call into text. Use a transcript if possible. If not, write a same-day summary while the details are fresh.
2. If the workflow starts from a recorded call, interview, coaching session, or other spoken source, Plaud can be a dedicated capture option. Note Pro is the better fit for desk meetings and conference-room calls where you want reliable room capture before drafting.
3. Pull any pre-call questionnaire, proposal notes, or email clarification into the same working set.
4. Convert odd formats into plain text, DOCX, or PDF.
5. Make one short file called engagement-rules.txt with your standard revision limit, payment terms, and what counts as out of scope.
6. Make another short file called client-context.txt with the client name, industry, stakeholders, and primary success outcome.
7. Add pricing-notes.txt only if pricing was already decided by you. Do not ask Claude to invent prices from the call.

Build the custom skill for this workflow

Create a folder named sow-from-discovery-call. Inside it, add a file named Skill.md. Paste the starter content below, then adjust it for your standard SOW sections, revision rules, acceptance language, and approval terms.

---
name: sow-from-discovery-call
description: Convert a discovery call transcript and intake notes into a statement of work with scope, milestones, assumptions, exclusions, and open questions.
---

# SOW From Discovery Call

Use this skill when the user wants to turn a discovery call transcript, intake form, or onboarding notes into a first-draft statement of work.

## What this skill should do

- Extract confirmed goals, scope, deliverables, timelines, dependencies, and approvals from the source material.
- Separate what is confirmed from what is assumed or still open.
- Draft a clean first-pass SOW with reusable headings.

## Required sections

- Project summary
- Scope
- Deliverables
- Milestones or timeline
- Assumptions
- Exclusions
- Open questions
- Approval or next-step language

## Guardrails

- Do not invent pricing, dates, or deliverables.
- Do not imply legal finality.
- Keep any unclear item visible.
- Treat this as a draft for human review before sending.

## Examples

- Turn this discovery call transcript into a first-draft SOW.
- Draft a statement of work from this intake form and call transcript.
- Use my SOW workflow and separate confirmed scope from assumptions.

Zip the sow-from-discovery-call folder so the ZIP contains the skill folder at the root. Upload the ZIP in Claude under Customize > Skills, then enable the skill. Do not upload a ZIP that contains loose files at the root without the skill folder.

This starter is a first draft, not a final contract template. If you always use the same milestone labels, exclusion language, revision policy, payment terms, or client approval wording, add those rules to Skill.md before you upload it.

Optional official Anthropic skill

Anthropic's official doc-coauthoring skill is the closest public fit for structured document drafting. Use it when you need general help shaping a document, improving a draft, or coauthoring a section-by-section deliverable.

Use the private sow-from-discovery-call skill when you repeatedly turn discovery calls into the same SOW shape, with your own scope rules, assumptions, exclusions, approval language, and open-question handling.

Why this should be a custom skill, not a one-off prompt

An SOW has a fixed structure and real risk. A custom skill keeps your standard sections, scope boundaries, assumptions, exclusions, and review rules in one reusable place. That reduces prompt rewrite friction and makes it harder for a brainstormed idea from the call to slip into the committed scope.

Skill trigger test

After enabling the skill, try these prompts with a small sample:

• Draft a statement of work from this discovery call transcript.
• Use my SOW workflow on this client intake form and onboarding notes.
• Turn this recorded discovery call into a first-pass scope document with open questions.
• Create a draft SOW and separate confirmed deliverables from assumptions.

Claude should treat these as sow-from-discovery-call tasks and produce the required sections with unclear items visible. If it does not, tighten the skill description so it names discovery calls, intake forms, SOW drafts, confirmed scope, assumptions, exclusions, and open questions.

Numbered workflow steps

1) Gather and clean the source material

Collect the transcript or call summary, intake form, proposal notes, email clarifications, client-context.txt, engagement-rules.txt, and any confirmed pricing notes. Remove duplicate transcript blocks, irrelevant small talk, and any material you do not have permission to process.

Create or update:

• client-context.txt for client name, industry, stakeholders, and success outcome
• engagement-rules.txt for revision limits, payment terms, approval steps, and standard exclusions
• call-source.txt for the transcript, cleaned call notes, or Plaud transcript export
• open-items.txt for anything you already know needs follow-up

2) Create, upload, and enable the custom skill

If you have not already done it, create the sow-from-discovery-call folder, add Skill.md, zip the folder, upload it in Claude under Customize > Skills, and enable it.

If the skill already exists, open Claude's Skills settings and confirm it is enabled before starting the SOW run.

3) Run a quick trigger test

Use one short prompt before uploading the full call package:

Use my SOW workflow to turn this into a first-pass scope document:

Client wants a website refresh. Confirmed: homepage rewrite and service page outline. Possible: new case study, not confirmed. Timeline discussed: June launch, exact review dates unclear.

Check that Claude separates confirmed scope from possible work, marks the review dates as unclear, and does not invent pricing, milestones, or deliverables.

4) Run the actual workflow with the skill active

Upload or paste the cleaned call material, client-context.txt, engagement-rules.txt, and any approved pricing notes. Then ask:

Use the sow-from-discovery-call skill to turn this discovery call material into a first-draft statement of work.

Use only the attached or pasted source material. Separate confirmed scope from assumptions and open questions. Do not invent pricing, dates, milestones, deliverables, or included work. Keep unclear items visible for my review before I send anything to the client.

If Claude asks for missing details, answer only with information you actually confirmed. If you are not sure, tell Claude to keep it in Open questions.

5) Do a human review for scope, pricing, and legal boundaries

Before sending, review the draft yourself or with your contract reviewer. Check:

• pricing and payment terms
• legal or liability language
• included and excluded work
• revision limits
• client responsibilities
• acceptance criteria
• milestones, dates, and dependencies

The skill should not decide these terms. It should preserve uncertainty so you can make the business decision.

6) Save, send, or reuse the result

After review, save the final SOW, source material, and any useful corrections in your client folder or Claude Project. If Claude missed a recurring boundary, update Skill.md before the next discovery-call workflow.

Tool-specific instructions

Claude Skills

• Use the page-specific sow-from-discovery-call custom skill first.
• Keep doc-coauthoring in mind as the closest official public skill for general structured document drafting, not as a full replacement for your private SOW rules.
• Anthropic's official skill-creator skill can help draft, test, or refine your custom skill if you want help improving the description, trigger prompts, or guardrails.
• For broader setup context, examples, and related Claude Skills workflows, see the Claude Skills tool hub.

Low-friction fallback

If you cannot upload a skill yet, keep the starter Skill.md text, one approved SOW example, client-context.txt, and engagement-rules.txt in a Claude Project. Paste the skill instructions before each run. This is less reliable than a real skill because you must remember the setup every time.

Quality checks

1. Every deliverable in the SOW can be pointed back to the transcript or a manual confirmation note.
2. Assumptions and exclusions are visible and not buried at the end.
3. Timeline language includes dependencies, reviews, or approval steps where those matter.
4. The document is specific enough that a new reader could tell what is included and what is not.
5. Pricing, legal terms, and approval language have been reviewed by a human before sending.

Common failure modes and fixes

The SOW commits to too much

Fix: update Skill.md to separate confirmed scope from assumptions and open questions. Rerun the trigger test before using the full transcript.

The tone sounds too salesy

Fix: add a tone rule to the skill. For example: "Use plain business language. Do not use persuasive proposal language inside the SOW."

The client can still misread what is included

Fix: add a deliverable acceptance line under each major deliverable. State what the client receives, in what format, and what counts as one revision.

Claude invents a milestone, price, or included deliverable

Fix: remove the invented item before sending. Add it to Open questions if it needs follow-up, then update the skill guardrails to keep unconfirmed dates, prices, milestones, and deliverables visible.