It’s time to get serious about ai.

Build Your Own AI Bidding Tool and Landing Page in 1 Weekend

PART ONE

SCHEDULE: Pacific Time

SATURDAY, MAY 8: 7AM TO 10 AM PART ONE
SATURDAY MAY 8: 1 AM – 4 PM PART TWO
SUNDAY: NOON TO 3 PM PART THREE

ZOOM LINK

PART ONE

PART  TWO

PART THREE

PART FOUR: PROMPTS

PART FIVE (WRAPUP)

How to use the prompts

Open ChatGPT (or Gemini, or Claude) and past the first prompt in. The AI will ask you questions, and you answer with as much detail as possible, but not more than one paragraph. When it is fiinished, it will create a file for you to copy and paste into a document of your choice: Word, Text… whatever.

Two prompts for you to complete before class on Saturday morning.

1. What are the twenty things you would want to know about me and my business before we wrote a business proposal or an email to a client? Format the above as a prompt I can answer.

2. I would like to create an overview of my business to add to my “about me” information for cold outreach emails, answering business queries, and formulating proposals to clients. Ask me 15 questions that will help you know exactly who I am and what I do for the basis of communications. Format this as a prompt I can answer.

The Kaparthy .MD you should always use when building anything.

# CLAUDE.md

Behavioral guidelines to reduce common LLM coding mistakes. Merge with project-specific instructions as needed.

**Tradeoff:** These guidelines bias toward caution over speed. For trivial tasks, use judgment.

## 1. Think Before Coding

**Don’t assume. Don’t hide confusion. Surface tradeoffs.**

Before implementing:
– State your assumptions explicitly. If uncertain, ask.
– If multiple interpretations exist, present them – don’t pick silently.
– If a simpler approach exists, say so. Push back when warranted.
– If something is unclear, stop. Name what’s confusing. Ask.

## 2. Simplicity First

**Minimum code that solves the problem. Nothing speculative.**

– No features beyond what was asked.
– No abstractions for single-use code.
– No “flexibility” or “configurability” that wasn’t requested.
– No error handling for impossible scenarios.
– If you write 200 lines and it could be 50, rewrite it.

Ask yourself: “Would a senior engineer say this is overcomplicated?” If yes, simplify.

## 3. Surgical Changes

**Touch only what you must. Clean up only your own mess.**

When editing existing code:
– Don’t “improve” adjacent code, comments, or formatting.
– Don’t refactor things that aren’t broken.
– Match existing style, even if you’d do it differently.
– If you notice unrelated dead code, mention it – don’t delete it.

When your changes create orphans:
– Remove imports/variables/functions that YOUR changes made unused.
– Don’t remove pre-existing dead code unless asked.

The test: Every changed line should trace directly to the user’s request.

## 4. Goal-Driven Execution

**Define success criteria. Loop until verified.**

Transform tasks into verifiable goals:
– “Add validation” → “Write tests for invalid inputs, then make them pass”
– “Fix the bug” → “Write a test that reproduces it, then make it pass”
– “Refactor X” → “Ensure tests pass before and after”

For multi-step tasks, state a brief plan:
“`
1. [Step] → verify: [check]
2. [Step] → verify: [check]
3. [Step] → verify: [check]
“`

Strong success criteria let you loop independently.
Weak criteria (“make it work”) require constant clarification.

—

**These guidelines are working if:** fewer unnecessary changes in diffs, fewer rewrites due to overcomplication, and clarifying questions come before implementation rather than after mistakes.

Meeting Recap

Quick recap

This was a technical training session led by Don where participants learned about AI tools and web development concepts. Don explained the difference between Claude desktop app (Claude Co-work) and the web version (Claude AI), emphasizing that the web version automatically saves work while the desktop version requires specific folder structures to maintain file accessibility. The class covered creating and managing prompts using Notion or similar tools, with Don recommending free platforms for organizing AI-generated content. Don demonstrated Notebook LM, a powerful research tool that can convert text into infographics, videos, and presentation slides, showing how it can process personal content like photography books, invoices, and client correspondence. The session included hands-on exercises where participants worked to condense their AI-generated About Me documents into 90-line versions and convert them to MD files for better processing. Don stressed the importance of being specific with prompts and using the right tools for different tasks, while also explaining how API keys work and the benefits of using self-hosted solutions like Netlify for deploying content.

Next steps

Don

Collaboration

Summary

AI Technology Class Planning Meeting

The meeting began with casual conversation between Luciana and Laura about weather conditions in Alabama and Laura’s recent move to a new house. The discussion then shifted to include Trish and Don joining the call, with Don expressing excitement about an upcoming class on technology. Don explained that he had originally planned an 8-week session but decided to adjust the timeline due to rapid developments in AI and imaging technology, aiming to ensure participants would be up to date with the latest information.

Web Development Fundamentals Overview

Don explained the fundamentals of web development, focusing on how content and structure are separated using databases, particularly in WordPress. He demonstrated how changing content in a database can update a website without modifying the page structure. Don clarified that while participants could use either the web version or Claude desktop app, they would work with the web version during the class for simplicity. The discussion included technical troubleshooting for participants using Claude, with Don addressing questions about API keys and Git commands. Don also mentioned introducing a new tool called Whisper Flow later in the class, which allows voice-to-text functionality across different platforms.

New Claude HTML Tool Demonstration

Don demonstrated a new tool he built with Claude that allows users to store information directly on HTML pages, showing how it can be used for idea management and other purposes. He explained the concept of APIs and warned about the security risks of exposing API keys, sharing his experience of having an API key stolen and used improperly. The team discussed their recent use of Claude prompts, with Luciana reporting she had generated 6-page documents for each prompt and questioning whether they needed to be more concise, while others reported receiving more concise responses with multiple format options.

AI Tools and Documentation Strategies

Don led a discussion on using different AI tools like ChatGPT, Claude, and Gemini, highlighting their unique approaches and features. He demonstrated how to create and customize an “About Me” document using ChatGPT’s personalization settings and shared a prompt template for generating a markdown file. Don advised participants to simplify their documents to 90 lines and convert them into markdown format for better organization. The group discussed using API keys for integrating AI tools with platforms like Substack for blogging, with Don recommending specific approaches for deploying content effectively.

Claude Desktop vs Web Issues

The meeting focused on troubleshooting issues with using Claude’s desktop application versus the web version. Don explained that the desktop version (Claude Co-work) requires specific file management practices, as moving files outside designated folders causes loss of access. Luciana learned that the web-based Claude (cloud.AI) automatically saves all work and conversations, eliminating the file management issues she experienced with the desktop version. The group clarified that the pro plan is required for using Claude Co-work, and participants were advised to use the web version for the current class activities to avoid confusion.

Prompt Organization Tools Comparison

Don discussed the differences between various tools for saving and organizing prompts, including Notion, Obsidian, and a desktop-based idea bank. He recommended Notion for its cloud-based accessibility and free, unlimited use, while noting that Obsidian can also be used with cloud storage. Don also demonstrated Notebook LM, a tool he uses to create infographics and videos, highlighting its cost-effectiveness and accessibility compared to hiring professional illustrators. He emphasized the tool’s ability to transform static documents into interactive content and its integration with other services like Gemini.

AI Content Creation Tools Demo

Don demonstrated various AI tools for content creation and editing, including voice conversion using 11 Labs, presentation slides in Canva, and Notebook LM for analyzing personal content. He showed how to convert audio to different voices, create editable slide decks from PDFs, and use Notebook LM to analyze photos, invoices, and client correspondence. Don also explained how to customize presentation styles with specific color palettes and design guidelines, and discussed the capabilities of different video formats including cinematic and explainer videos.

Notebook LM Capabilities Demo

Don demonstrated Notebook LM’s capabilities, showing how it can analyze and generate insights from various content types including images, PDFs, and documents. The tool can create different formats like mind maps, study guides, infographics, and data tables, with the paid version offering additional features. Don explained that users can integrate external sources like ChatGPT through a third-party tool called Cortex, which costs $3.75 per month or includes free access with Google Workspace. The discussion revealed that Notebook LM can automatically pull information from Google services like Gemini, making it easier to import existing content without manual copying and pasting.

Business Proposal Tool Development

The team discussed building a bid tool to help with creating business proposals and contracts. Debra, Mike, and Luciana expressed interest in developing a tool that would include contract terms and allow for electronic signatures. Don clarified that while the tool could generate a visually appealing proposal, it would not serve as the actual contract due to technical limitations. The group agreed to focus on creating a simplified proposal tool rather than incorporating complex features like image libraries or direct CRM integration. Don explained that building such a tool would require server setup and API integration, which was beyond the scope of the current project.

Client Proposal Process Discussion

The team discussed creating and sending proposals to clients, clarifying the difference between proposals and bids. They agreed that the current proposal tool could serve as a cosmetic presentation to clients but would need to be followed by a formal contract and invoice. Don demonstrated how the proposal could be converted into a clean PDF. The team decided to meet again at 1 o’clock to work with Claude to define how the tool should function and address remaining questions about the process.

Customizable Proposal Tool Development

The team discussed building a customizable tool for creating proposals, addressing whether it should be a single tool with customizable parameters or separate tools for different niches. Don clarified that while niche-specific tools like architecture might be needed for certain specialized cases, a general tool would work for most photography needs including food, products, and location work. The group then began listing required fields for the tool, including client information, project details, deliverables, and usage rights, with agreement that the tool would need to be customizable for different clients and project types.

Client Proposal Template Development

The team discussed creating a comprehensive proposal template for client projects, focusing on various deliverable options including print, film, and video content. They outlined key considerations for project planning, including timeframes, styling needs, props, location requirements, talent requirements, and insurance needs. Don shared a personal anecdote about a high-value jewelry shoot that required armed security, emphasizing the importance of proper insurance coverage for valuable items. The team agreed to include detailed breakdowns of costs as leverage when discussing pricing with clients.

Photography Bid Template Development

The team discussed creating a comprehensive bid template for photography services, focusing on various aspects including equipment rental, post-production, travel fees, and client information requirements. Don led the discussion on pricing structure, emphasizing the importance of including a flat fee for gear rental and building travel costs into the creative fee rather than listing them separately. The group agreed to add client contact information, project ID, and deliverable details to the template. They planned to reconvene in 3 hours to further develop the tool, with Don suggesting they consider design elements like fonts and colors for the final product.

This will be our home page for the workshop.

Build Your Own AI Bidding Tool and Landing Page in 1 Weekend

And we will do it.

You will need ZOOM for the meetings.

You will also need the Pro version of Claude ($20 per month) in order to get an API key. You can buy a small amount of API calls. For this class, $5 will be more than enough. (they are 1-cent per call)

Bookmark this page:

And watch your email and this page for more info as we get ready to launch the course.

___________________________________________________________________________________________

How to make Claude work for YOU…

1. Give Claude your actual context, not a generic request

“Write me a client email” gets you a generic email. “Write me a follow-up email to a bride who hasn’t paid her final invoice, due three weeks ago, in a tone that’s warm but firm because I want to keep the relationship good for referrals” gets you something you can actually send. The more specific the situation, the more useful the output.

Weak Prompt:

Write a follow-up email to a client about an unpaid invoice.”

Strong Prompt

“Write a follow-up email to a bride named Sarah. Her final invoice of $800 has been due for three weeks. She was a great client, the wedding went well, and I’ve already sent one reminder. I want to be warm but direct because she’s likely to refer friends to me. Don’t make it sound like a threat, but I do need payment this week.”

Weak Response

Subject: Invoice Follow-Up

Hi,

I hope you’re doing well. I wanted to follow up regarding the outstanding invoice for your recent photography services. Could you please let me know when we can expect payment?

Thank you for your time, and I look forward to hearing from you.

Best regards, [Your Name]

Stronger Response

Subject: Quick Note on Your Final Invoice

Hi Sarah,

I hope you’re still riding the high from your wedding day. It was such a beautiful event and I loved being part of it.

I wanted to reach out about the final invoice of $800, which has been outstanding for a few weeks now. I know post-wedding life gets busy fast, so I wanted to check in directly rather than let it slip further.

If you could arrange payment by Friday, I’d really appreciate it. I accept [payment methods]. If there’s something going on that I should know about, I’m happy to have a quick conversation.

Thanks so much, Sarah. Wishing you and [partner’s name] a wonderful start to married life.

[Your Name]

2. Treat it like a first draft machine, not a finished product machine

Claude’s best role in a photography business is producing a strong 80% draft fast, which you then make yours. Contract clauses, gallery delivery emails, Instagram captions, FAQ pages for your website. The goal is to cut the blank-page problem, not to copy-paste without reading.

3. Reuse what works with saved prompts

Once a photographer finds a prompt that produces a great output (say, a specific way to describe their editing style for inquiries), they should save that prompt. You’re essentially building a personal toolkit over time. Show them a simple example of a prompt they’d use weekly.

4. Upload your actual documents and have Claude work with them

Photographers can paste in their current pricing guide, questionnaire, or contract and ask Claude to improve the language, find confusing sections, or rewrite it for a different client type (budget vs. luxury). Working from real material beats building from scratch every time.

5. Use it to prep for difficult conversations

This one surprises people. Photographers can describe a tricky client situation and ask Claude to help them think through how to respond, what they might be missing, or how to phrase a boundary clearly. “My client wants to reschedule their session for the third time, two days before it happens. Here’s my contract language. How should I respond?” It functions well as a thinking partner before a stressful call or email.

Claude works best when you treat it as a capable assistant who needs the full picture, not a vending machine you drop a vague question into.

The “About Me” file, and where to put it.

1. Claude.ai Custom Instructions (Best starting point)

Inside Claude.ai, under Settings, there’s a section called “Personal Preferences” where they can write a standing description of themselves and their business. Claude reads this at the start of every conversation automatically.

You should put things like:

  • Full name and business name
  • Location and the type of work they specialize in (weddings, commercial, portraits, etc.)
  • Your typical client (luxury brides, small businesses, families)
  • Tone you write in (warm and personal, clean and professional, etc.)
  • Any specific language you always or never use
  • Website URL, email, and social handles if relevant

Once it’s there, they don’t explain themselves every conversation. Claude already knows who they are.

2. A Personal “About Me” Block They Paste When Needed

For more detailed or specific work, you should keep a short text block saved somewhere easy (Notes app, a Google Doc, their phone) that they paste at the top of any prompt. Something like:

“Here’s my background for context: I’m [Name], owner of [Business]. I photograph [specialty] in [location]. My clients are typically [description]. My writing tone is [description]. My website is [url].”

This is useful when the task is involved enough that you want Claude fully loaded with context before it writes anything.

The best plan: Both.
Settings handles the everyday stuff automatically.
The paste block is there when the stakes are higher and you want Claude working with the full picture.

Here’s mine:

  • Don Giannatti, Don Giannatti Studios, Phoenix, Arizona:
  • Specializing in tabletop, product, and small business website design
  • My typical client is a small business that needs to increase their visibility with better images and a website that gets attention.
  • I write in a warm, friendly tone, but with a no-nonsense approach
  • Stay away from corporate speak, and typical AI phrasing.
  • [www.dongiannatti.com] [www.dongiannattiphotography.com]/ @wizwow

JSON isn’t necessary here. Claude reads plain text just fine, and converting this to JSON adds complexity with zero benefit for this use case. JSON is useful when software needs to parse structured data. Claude just needs to read and understand it.

Here’s what Claude created:

Don Giannatti | Don Giannatti Studios Phoenix, Arizona

I specialize in tabletop and product photography, and website design for small businesses. My clients are typically small business owners who need better images and a website that actually gets noticed.

My tone is warm and friendly, but direct. No fluff, no corporate speak, no filler phrases. Write like a real person talking to another real person.

Web: dongiannatti.com / dongiannattiphotography.com Social: @wizwow