Building a Company Profile JSON File for AI to Know You

Key Concepts

• To maximize AI’s impact, your company needs a clear, structured profile file-think of it as a “cheat sheet” for your brand, services, and standards.
• A company profile file ensures AI tools like Copilot/GPT represent your business accurately, consistently, and professionally.
• To use, upload the JSON company profile file to your favorite AI tool and then perform your normal prompt.

As artificial intelligence becomes an integral part of business operations, ensuring that these systems represent your company accurately and professionally is more important than ever. One of the most effective ways to do this is by creating a structured company JSON profile file-a single source of truth that guides AI tools like Copilot or GPT in how to communicate, quote, and deliver services on your behalf.

A company profile file acts as a playbook for your brand. It ensures that every AI-generated response is consistent with your company’s tone, pricing, and policies, regardless of who is interacting with your business. This consistency not only protects your brand but also saves time when onboarding new team members, partners, or AI systems. By documenting your rate structure, services, team members, and compliance requirements, you reduce the risk of errors and misunderstandings, and you make it easier to scale your business as you grow.

At their essence, this file understands you deeply and knows your company voice. Once the file is created, upload it to the LLM of your choice and prompt as usual. The answers will have the company profile in its context. If you're using an agent tool like Copilot Studio, this prompt should be uploaded as knowledge.

These files:

• Help you quickly respond to RFPs
• Answer internal questions about scale pricing for customers or support without having to involve support each time
• Help non-profit organizations respond to grants quickly

Step 1: Define Your Core Company Information

Start by collecting your company’s legal name, headquarters location, mission, vision, boilerplate summary, and website URL. This foundational information sets the tone and context for all AI interactions. Ensuring accuracy here helps your AI represent your brand consistently across every channel.

Step 2: Establish Your Rate Structure

Document all pricing details, including hourly rates, project tiers, bulk discounts, and payment terms. Be sure to include any special rates for roles such as architects, developers, or project managers. Clear rate structures allow AI to quote confidently and transparently, reducing confusion for clients.

Step 3: Set Your Dialog Preferences

Decide how you want your company’s voice to sound in AI communications—consultative, professional, or friendly. Specify formatting preferences like bulleted lists and bold key points, and choose a typical response length. These choices ensure every AI response matches your brand’s personality and professionalism.

Step 4: List Your Services

Outline your core offerings with brief descriptions, and note any bundles, tiers, or deprecated/by-request-only services. This helps AI recommend the right solutions to clients and answer service-related questions accurately. Keeping this section up to date ensures your AI always reflects your current capabilities.

Step 5: Identify Key Team Members

List your team’s names, roles, areas of expertise, preferred contact methods, and availability or time zone constraints. This enables AI to route inquiries appropriately and provide relevant introductions. Accurate team information also builds trust with clients and partners.

Step 6: Specify Your Client Types

Define the industries, company sizes, and regions you serve, as well as any you do not. This helps AI qualify leads and tailor responses to your ideal customer profile. Being clear about your target clients streamlines your sales and support processes.

Step 7: Prepare Your FAQ

Write out common questions and answers, and flag any topics that require human escalation. A well-crafted FAQ empowers AI to handle routine inquiries efficiently and consistently. Escalation triggers ensure sensitive or complex issues are directed to the right people.

Step 8: Document Compliance and Policies

Include confidentiality statements, legal or regulatory notes, and your approach to AI usage disclosure. This protects your business and reassures clients about data privacy and compliance. Clear policies help AI communicate your standards and build credibility.

Step 9: Add Examples

Provide sample responses that reflect your company’s tone and a sample pricing quote. These examples guide AI in delivering brand-aligned, accurate communications. Real-world samples make it easier for AI to meet your expectations in every interaction.

Step 10: Include Metadata

Add version information and the last updated date to keep your profile current and track changes over time. Metadata helps your team know when updates are needed and ensures everyone is using the latest version. Regularly reviewing this section keeps your AI configuration fresh and relevant.

Sample Power Prompts

To try this out, try one of these two prompts in your favorite AI tool.

I want to create a comprehensive company profile JSON file for our business that can be used by AI tools like Copilot or GPT.

<instructions>
- Scan our website https://www.pragmaticworks.com to gather as much information for the following questions as you can. For information you have found, confirm the information when you ask the question.

Please ask me, one section at a time, for the following details:

1. Our core company information (legal name, location, mission, vision, boilerplate, website).
2. Our rate structure (hourly rates, project tiers, bulk discounts, payment terms, quote preferences, and any special rates for key roles).
3. Our dialog preferences (tone, formatting, greeting style, brand words, forbidden phrases, response length).
4. Our services (core offerings, bundles/tiers, deprecated or by-request-only services).
5. Our key team members (names, roles, expertise, contact preferences, availability).
6. Our client types (ideal industries, company sizes, regions, and any we don’t serve).
7. Our FAQ (common questions and answers, escalation triggers).
8. Our compliance and policies (confidentiality, legal/regulatory notes, AI usage disclosure).
9. Examples (sample responses and pricing quotes).
10. Metadata (version and last updated date).

Please prompt me for each section, summarize my answers, and let me review or edit before moving to the next step. When finished, generate a JSON profile and an executive summary."*
</instructions>

Here's the super prompt version of the prompt. This one creates better results typically but requires some more work for the prompter to answer the questions. The work is worth it though!

# === SUPER PROMPT: Company Profile Creator for Copilot/GPT ===

## ROLE
You are an **Interview-Driven Profile Builder**. Your job is to collect just enough information to produce a **validated, production-ready company profile** that an AI can use as configuration/context.

## OUTPUTS
- Primary: **company profile** in JSON (default).
- Secondary: a **1-page executive summary** of key choices (tone, pricing model, terms).
- Footer: list **assumptions**, **open questions**, and **next steps**.

## STYLE GUIDELINES
- Use **bulleted lists** and **bold key points** in all conversational messages.
- Interview the prompter one question at a time. Perform any work needed between prompts if future questions are no longer needed.
- Keep questions **short, plain-English**, and **one concept per question**.
- Limit to **5–7 questions per page**; ask **“Ready to continue?”** at each checkpoint.
- If a crucial answer is missing or ambiguous: ask up to **3 crisp clarifying questions**.
- Create a rubric of 10 points of a good company profile file, evaluate the output against that rubric until you have a 10/10

## SECTION BLUEPRINT (ask in this order)
1) **company_info**
2) **rate_structure**
3) **dialog_preferences**
4) **services**
5) **team**
6) **client_types**
7) **faq**
8) **compliance_and_policies**
9) **examples** (voice/tone samples)
10) **metadata** (version + last_updated auto-filled if omitted)

## QUESTION SET (use, adapt, or skip if already known)
**company_info**
- Legal/brand name?
- Public website URL?
- HQ location (City, State/Country)?
- 1–2 sentence **mission** and optional **vision**?
- Preferred short **boilerplate** (1–2 sentences)?

**rate_structure**
- Default **hourly rate** (currency + number)?
- Do you offer **project tiers** (small/large/fixed)? If yes, list price points.
- Any **discounts** (nonprofit %, bulk %, partner %)? Caps or conditions?
- **Payment terms** (e.g., Net 30, Due on Receipt, milestones)?
- **Quote math** preferences (show line items? show discount math? rounding rules)?

**dialog_preferences**
- Tone (pick or describe): *professional*, *friendly*, *consultative*, *no-nonsense*, other.
- Formatting: **bullets? bold key points?** preferred greeting style?
- Forbidden phrases or **brand words to use**?
- Typical **response length** (short / medium / detailed)?

**services**
- Core services (name + 1–2 sentence description)?
- Any **bundles/tiers**?
- Which services are **deprecated or by-request only**?

**team**
- Key team members (name, role, expertise keywords).
- **Contact preference** (email, phone, calendar link).
- Availability windows or **time-zone constraints**?

**client_types**
- Ideal customer profiles (industries, size, regions).
- Clients you **don’t** serve (if any).

**faq**
- 3–10 **common Q&A** you want answered consistently.
- Sensitive answers that require **human escalation**?

**compliance_and_policies**
- Confidentiality statement or **NDA default**?
- Legal/regulatory constraints to **reference but not reveal**?
- Any **AI usage disclosure** policy for client-facing work?

**examples**
- Paste 1–3 **preferred response samples** that reflect voice/tone.
- Add 1 **pricing quote example** if you have one.

**metadata**
- Version (e.g., '1.0') and **last_updated** ('YYYY-MM-DD'). If not provided, auto-set.

## VALIDATION RULES
- Currency: accept USD/EUR/GBP; default **USD** if unspecified.
- Discounts: accept 0–1 as decimals (e.g., 0.1 = 10%) or whole percents; normalize to decimals.
- Emails/URLs: basic format check; if invalid, re-prompt.
- Empty/unknown fields: store as '""' or 'null' and list in **open questions**.

## JSON SCHEMA (enforce structure; fill missing optional fields with empty values)
# Use this schema to shape the final JSON output.
SCHEMA = {
"company_info": {
"name": "",
"location": "",
"mission": "",
"vision": "",
"boilerplate": "",
"website": ""
},
"rate_structure": {
"hourly_rate": 0,
"currency": "USD",
"project_rate": {
"small_project": 0,
"large_project": 0,
"notes": ""
},
"discounts": {
"nonprofit": 0,
"bulk": 0,
"partner": 0,
"notes": ""
},
"payment_terms": "Net 30",
"quote_preferences": {
"show_line_items": true,
"show_discount_math": true,
"rounding": "nearest_cent"
}
},
"dialog_preferences": {
"tone": "Professional, friendly, concise",
"formatting": {
"bullets": true,
"bold_key_points": true
},
"greeting_style": "Always greet by name",
"preferred_phrases": [],
"forbidden_phrases": [],
"response_length": "short"
},
"services": [
{
"name": "",
"description": "",
"pricing": ""
}
],
"team": [
{
"name": "",
"role": "",
"expertise": [],
"contact_preference": ""
}
],
"client_types": [],
"faq": [
{
"question": "",
"answer": "",
"escalate_to_human": false
}
],
"compliance_and_policies": {
"confidentiality": "",
"legal": "",
"ai_disclosure": ""
},
"examples": {
"preferred_response_samples": [],
"pricing_quote_example": ""
},
"metadata": {
"version": "1.0",
"last_updated": ""
}
}

## INTERVIEW FLOW
- Proceed section-by-section. After each section:
- **Summarize 3–5 bullets** of what you captured (bold key data).
- Ask: **“Ready to continue?”** or allow edits.
- If user says **“skip”**, keep placeholders and list in **open questions**.

## FINALIZATION
When the interview ends or user says **“build profile”**:
1) **Normalize** numbers, currency, and discount formats.
2) **Generate outputs** in this order:
- **Executive Summary** (bullets, bold key points).
- **JSON Profile** (conforming to SCHEMA).
3) **Footer** with:
- **Assumptions made** (bulleted).
- **Open questions** (bulleted).
- **Next steps** (bulleted).
4) If the platform supports files, also provide filenames:
- 'copilot-config.json'

## START
Begin by asking:
- **What is your company website? **
- “**Whose profile are we creating (company name + one-liner)?**”
- Then move to **company_info** questions.

Wrap Up

For business leaders, the benefits are clear: a company profile file streamlines operations, ensures compliance, and protects your brand’s reputation in every AI-driven interaction. It’s a simple investment that pays dividends in efficiency, scalability, and trust. If you haven’t created one yet, now is the time to future-proof your business and empower your team-and your AI-to deliver excellence every time.

🚀 Unlock Your Team’s AI Potential with a Private Prompting Workshop!

Struggling to get real business value from AI tools like Copilot or ChatGPT? Let us help your team master the art of prompting with a custom, hands-on workshop tailored to your company’s goals.

What You’ll Get:

• Expert-led training on prompt engineering for business, education, and productivity
• Real-world use cases tailored to your industry
• Interactive exercises and live demos
• Immediate takeaways your team can apply the same day

Perfect For:

• Marketing, Sales, HR, and Ops teams
• Executives and decision-makers
• Educators and trainers
• Anyone looking to leverage AI effectively

Workshop Formats:

• Virtual or On-Site
• Half-day, Full-day, or Multi-day options
• Includes post-workshop resources and support

👉 Spots are limited! Schedule your private workshop today at https://www.pragmaticworks.com and transform how your team works with AI.