As a solo creator managing a blog, Pinterest, and an Etsy shop, I handled everything from identifying opportunities to publishing content and reviewing performance. I already used generative AI, but repeated context-setting and manual handoffs made production time-consuming, leaving little time for analysis and strategic decisions.

I managed the entire content lifecycle across multiple channels.

THE PROBLEM

One creator, many responsibilities

A lot to manage, but no unified system

Fragmented AI workflow

Too much time on operations

Each new task required me to repeat business context, audience details, and brand guidelines. I also moved information manually between tools and corrected inconsistent outputs.

As channels grew, fragmented production left little time to analyze performance and prioritize next steps.

DESIGN GOALS

A more strategic and efficient process

The goal was to create a more connected and efficient content operation, reducing repetitive work while maintaining creative control and consistency across channels.

To achieve this, I defined six priorities:

Identify opportunities

Plan with SEO in mind

Create more efficiently

Reduce the need to repeatedly explain the same business context to general-purpose AI tools.

Maintain human review and approval throughout the process.

Organize performance data to support future decisions.

Identify relevant content and product opportunities to prioritize.

Consider keywords and search intent before drafting content.

Reduce production time for articles, images, Pinterest assets, and Etsy listings.

Retain business context

Stay in control

Learn from data

Mapping the Previous Workflow

Before creating the system, I mapped the main stages involved in managing my content ecosystem:

The lack of a connected process created rework, inconsistent outputs, and unnecessary context switching.

Although I already used ChatGPT during some of these stages, it was not customized to my business. I frequently had to reintroduce my audience, content categories, products, preferred writing style, SEO requirements, and visual direction.

PREVIOUS PROCESS

Designing the System

An overview of how specialized AI assistants, automation workflows, and human decisions work together across the content lifecycle.

THE SOLUTION

I started with image generation, the most time-consuming part of my workflow, before expanding the system to other stages of content production.

Instead of relying on a single general-purpose assistant, I designed the system as a set of specialized components, each with its own responsibility, business context, inputs, and outputs.

This modular approach allows me to refine each component independently. Custom GPTs and AI-powered n8n workflows support research, content creation, and recommendations. n8n also handles repeatable tasks, data collection, and connections to Google Sheets, where I organize outputs, review information, and track progress.

Custom GPTs use their configured instructions and reference materials rather than accessing Google Sheets directly. I coordinate the handoffs between components and review opportunities, creative outputs, and publishing decisions.

The System at a Glance

From scattered signals to structured opportunities

I developed a Market Scout assistant that analyzes the market to identify emerging or underserved opportunities for content and digital products. It evaluates multiple signals using consistent criteria and structures the findings for my review and decision-making.

Market Scout uses Web Search to research current public information. My playbook guides the research toward sources such as Etsy, Pinterest, Google Trends, and relevant cultural publications or event calendars. The model evaluates these signals against my business context and records reference links alongside the findings.

DISCOVERY

Key evaluation criteria

• Target audience and geographic relevance

• Seasonality and expected demand

• Competition and differentiation opportunities

• Production effort

• Intellectual property risk

• Potential across Etsy, Pinterest, and the blog

• Recommended timing and priority

01

02

03

Request & context

AI analysis

Structured output

The workflow starts when I send a chat message with a topic or request, along with any specific context.

The model analyzes the request using my business context and evaluation criteria, including audience, seasonality, demand, competition, and production effort.

The workflow prepares opportunities, deep dives, and references, and saves them in organized sheets for easy review.

I translated these requirements into a weighted scoring framework in the Market Scout playbook. The AI model uses this framework to assess opportunities from 0 to 100, while evaluating intellectual property risk separately.

Here’s how the Market Scout moves from an initial request to structured opportunities.

Opportunities organized in a spreadsheet

After noticing content on TikTok around women’s themed gatherings, I asked Market Scout through the n8n chat to evaluate opportunities for watch-party printables, girls’ nights, and related occasions in the U.S. market.

One recommendation was a Creative Social Night Host Kit, an editable kit to help hosts organize a creative gathering. It received an opportunity score of 83/100. The Opportunities sheet lets me compare this direction with other ideas by audience, product format, score, and recommendation.

Turning criteria into an evaluation framework

n8n saves the results across three linked sheets: Opportunities for comparison, Deep Dives for detailed analysis, and References for supporting sources. I use this information to decide what to create, investigate further, or set aside.

Exploring an initial signal

Understanding the context behind the score

Because my cultural and market references come primarily from Brazil, I need more context to evaluate opportunities in the U.S. The Deep Dives sheet explains relevant customs, market signals, competitor patterns, and potential gaps, helping me assess whether an idea fits my product style and production capacity.

For this opportunity, the analysis suggested moving beyond standalone invitations toward a coordinated host kit that supports the full gathering, from setting up activity stations to guiding the experience.

Research commentary is generated in Brazilian Portuguese for my review, while U.S. market terms and product names remain in English.

Reviewing supporting references

The References sheet links each opportunity to source URLs and market examples. This gives me a way to explore the cultural context, inspect existing products, and review the evidence behind a recommendation before deciding what to create.

I created a specialized content workflow that incorporates SEO from the beginning, rather than treating optimization as a final checklist.
This helps me define the topic, structure the article, and plan the visuals before writing, resulting in higher-quality content in less time.

• Search intent and target keywords.

• Article structure.

• Related questions and supporting topics.

• Internal content and product opportunities.

• Image requirements for each section.

• Image filenames and alt text.

• Suggestions for creating, photographing, or sourcing each visual.

CREATE - ARTICLE

From idea to a structured, SEO-optimized article

What this step covers

01

Explore the topic

Discuss the idea, target audience, and search intent.

02

Generate structure

Get a detailed outline with relevant subtopics and questions.

03

Plan the visuals

Define image ideas, filenames, and alt text for each section.

04

Write and refine

Draft the article and iterate with my feedback.

05

Review and polish

Finalize the content and manually transfer it to WordPress.

Custom GPT configuration

Defined role, instructions, and business context to keep the output aligned with my goals and brand voice.

Chat interface

For a watch-party article, I provided the topic and product context. The assistant proposed a brief for my review before outlining and drafting. Planning takes place in Portuguese; final articles are written in U.S. English.

Here’s how the new workflow moves from an initial idea to a structured, polished article ready for publishing.

I created a specialized image assistant and a visual playbook documenting my brand guidelines and image creation requirements. The assistant uses this reference to turn content needs into visual concepts and prompts across channels. I review the results, request refinements when needed, and decide whether each image fits the brand.

• Translate content goals into image concepts.

• Maintain more consistent art direction.

• Produce prompts for different formats.

• Plan SEO information for website images.

• Generate Pinterest-oriented variations.

• Reduce repeated prompt refinement.

CREATE - IMAGE DIRECTION AND GENERATION

From idea to on-brand visuals

What the assistant supports

01

Define the visual direction

Turn the content need into visual direction, concepts, and prompts.

02

Generate and refine

Create the image and iterate on composition, style, or details.

03

Finalize in Canva

Apply brand elements, text, and templates when needed.

04

Adapt across channels

Use the final visuals across the blog, Pinterest, and Etsy.

Here’s how that direction moves from the specialized assistant to finished visuals across my channels.

Pinterest

Vertical images with clear messaging, designed for Pinterest discovery and engagement.

Blog

Featured and in-content images to make articles more visual and engaging.

Etsy

Lifestyle mockups and product images to showcase the printables in real settings.

I built an n8n workflow that uses product images, a short brief, and a listing playbook I created to generate an Etsy draft. It fills in the listing fields, leaving the images and digital files for me to upload before reviewing and publishing.

• Listing titles and descriptions.

• Keyword and tag suggestions.

• Product information and customer instructions.

CREATE - LISTING CREATION

From inputs to a ready-to-edit listing

What this step covers

01

Provide the inputs

Upload two product images and a short brief.

02

Generate the listing

AI uses my listing playbook and product inputs to generate structured content and populate the listing fields.

03

Create the Etsy draft

The workflow sends the listing to Etsy as a draft.

04

Review and publish

I upload the product images and digital files, review the listing, and publish it.

Here’s how the workflow turns product inputs into a structured Etsy draft, ready for review and publishing.

I built an n8n workflow that turns a blog article or Etsy product into a structured Pinterest campaign, generating multiple creative directions, copy, and visual guidance from the source content.

• Multiple pin concepts for the same destination

• Pin titles and descriptions

• Keywords and search-oriented angles

• Visual concepts and image direction

• Campaign structure and organization

• A clear connection between each pin and its source content

DISTRIBUTE - PINTEREST CAMPAIGN PLANNING

Turn one piece of content into a Pinterest campaign

What the workflow generates

01

Provide the source

Add the URL of a blog article or Etsy product and define how many pin ideas to generate.

02

Generate the campaign

n8n uses AI to create pin concepts, titles, descriptions, keywords, and visual direction.

03

Review and organize

The generated content is saved to a spreadsheet. I check that the output is complete and correct, occasionally adjust titles, and create the images separately.

The first workflow focuses on campaign planning, turning the source content into multiple Pinterest-ready directions for review.

DISTRIBUTE - PINTEREST CAMPAIGN PUBLISHING

Testing Automated Pinterest Publishing

I built an n8n workflow to automatically publish approved Pinterest pins and update their status in Google Sheets.

The integration I tested published pins immediately, rather than saving them in Pinterest’s native scheduler for a final visual check. It also did not let me select the predefined topic tags available in Pinterest’s interface, which I wanted to retain for discovery.

04

Approve content

Test automated publishing

05

Use n8n to publish approved pins through the Pinterest API.

06

Validate the workflow

Confirm that pins were published and their status was updated in Google Sheets.

Review and approve the pins prepared in the campaign spreadsheet.

This experiment helped me define where automation was useful in my workflow. I kept campaign planning and content preparation automated, while handling final review and scheduling in Pinterest’s interface. Adding another scheduling tool would have introduced an extra platform without meaningfully reducing the remaining manual work.

TESTED APPROACH · NO LONGER USED

Collecting data automatically

I built an n8n workflow to collect key performance data from Pinterest, Etsy, and my blog (Google Analytics) weekly.

Google Sheets maintains a performance history organized by platform, collection date, and reporting period, helping me track changes over time and compare trends across channels.

MEASURE - DATA COLLECTION

01

Identify the key metrics from Pinterest, Etsy, and Google Analytics that can be combined to evaluate performance and identify patterns.

Select the data that matters

02

Build the data collection workflow

Create an n8n workflow that connects to each platform, collects the selected metrics, and sends them to a centralized Google Sheets dataset.

03

Centralize performance data

Consolidate the collected data in Google Sheets, manually adding Etsy metrics that are not available through the API.

Pinterest

• Account analytics

• Pin performance

• Outbound clicks

• Impressions, engagements and saves

Etsy

• Listing performance

• Views, favorites, sales

• Revenue and conversion

• Top products

Google Analytics (Blog)

• Page performance

• Traffic sources

• Sessions and engagement

• Top articles

LEARN AND DECIDE - ANALYSIS AND ACTION

From data to insights to action

I use a Growth & Priority Advisor GPT to explore performance data organized in Google Sheets. Through follow-up questions, I add business context, challenge initial interpretations, and develop hypotheses to test.

01

Identify a performance signal

I created the printable bundle and listing images, printed the items, and recorded a product video. I used the Etsy workflow to prepare the listing and published new Pinterest creatives weekly leading up to the premiere.

03

02

Connect research with context

Market Scout surfaced hockey romance as a potential theme. Connecting this with activity I had noticed around book adaptations, I searched TikTok and found examples of friends hosting small watch parties. I then considered the upcoming release of The Love Hypothesis.

Define a focused product test

With about three weeks until the premiere, I searched Etsy and found limited relevant options. Through discussion with the AI assistant, I narrowed the idea to a small printable party bundle that I could create and promote before the release.

Example: Testing a printable watch-party kit

A review of my World Cup party content revealed Pinterest visibility but few sales of the related printable products. This prompted me to explore a more specific buying occasion, eventually leading to a watch-party kit for friends gathering to watch the film adaptation of The Love Hypothesis.

05

Create and distribute

Reviewing three months of data with the AI assistant, I noticed that visibility on Pinterest had not translated into many World Cup printable sales. I began exploring whether a more specific occasion and less crowded niche could offer a stronger opportunity.

04

06

Observe the results

Learn and decide

The bundle generated six sales from the Friday before the premiere through release day, followed by one more sale the next week. Pinterest impressions, outbound clicks, and Etsy listing visits increased around the release, then declined afterward. I could not attribute individual sales to Pinterest.

The test provided an initial signal of buying interest in a specific watch-party occasion. The decline after the premiere suggested a short demand window. These findings gave me a basis for exploring similar release-driven opportunities, without assuming the same results would repeat.

This opportunity emerged through a combination of performance analysis, AI-supported research, my own observations, and follow-up questions. The system helped me turn those inputs into a focused product test and review what happened afterward.

Less operational work, more room for strategic decisions

Beyond production speed, the system helped standardize how I evaluate opportunities, prepare content, and review performance.

The biggest improvement was not simply producing more content. It was creating a repeatable process that allows me to spend less time on repetitive tasks and more time deciding what is worth creating next.

RESULTS & IMPACT

The system reduced the time required to prepare content across multiple channels while creating a more structured process for planning, production, and performance analysis.

ESTIMATED PRODUCTION TIME IMPROVEMENTS

~5h

Article production and publishing

Previously ~16h

~3.5h

14 Pinterest pins, created and published

Previously ~8h

~20min

Etsy listing creation and publishing, including images

Previously ~3h

Estimated task times based on my experience over six months with the previous workflow and two months with the new system. These are approximate estimates, not formally tracked measurements.

More structured planning

Market research, SEO planning, and content prioritization now follow more consistent criteria, reducing disconnected decisions.

More consistent production

Playbooks define content structures and visual guidelines. Pinterest outputs now follow a standard format with titles, descriptions, tags, links, and image concepts, making them easier to review and prepare for publishing.

Better visibility for decisions

Centralized performance data makes it easier to review results, identify patterns, and decide which opportunities to explore next.

Operational Improvements

API usage cost: $1.16 across 41 requests in September 2026.
This covers API usage only; subscriptions, hosting, and maintenance time are excluded.

From operational efficiency to continuous experimentation

NEXT ITERATION

What I’m exploring next

Improve the system

Refine AI assistants, simplify workflows, and identify where automation can reduce effort without compromising quality or control.

Increase qualified traffic

Validate product opportunities

Test content topics, SEO opportunities, and Pinterest creative strategies to understand which approaches attract more relevant visitors.

Use Etsy performance, market signals, and customer behavior to improve existing listings and test new product ideas.

With the core workflows in place, my focus is shifting from improving production efficiency to testing how the system can support better content and product decisions, while keeping the architecture flexible enough to adapt as AI capabilities evolve.

Designing for change, not just today's tools

My goal is to build an adaptable system that helps me learn from performance, refine workflows, and make better content and product decisions as the technology evolves.

As my custom GPTs transition to plugins, I plan to adapt their instructions and playbooks into reusable capabilities while preserving each component’s role in the workflow.
My next step is to evaluate this migration, validate the quality of each assistant's outputs, and preserve the separation between AI reasoning, automation, shared data, and human decisions.

Carla Oliveira

Home | About | Work | Contact

Fort Worth - TX - United States