Node Filter Unleashed: A Journey of Predictive Project Management

In the picture below, you can see my way of estimating how long tasks will take by using past information from my company’s graph. One of my colleagues nicknamed it “Node Filter Galore”.

In the busy environment of my digital design studio, handling multiple projects at the same time is common. My team enjoys the creative chaos, but as the project manager, it’s my job to ensure these projects are completed on time. My goal has always been to deliver excellent designs within the agreed timelines, a goal that requires careful planning of resources, skills, and schedules.

The process of estimating project timelines has traditionally been based on our experience and understanding of the tasks. We had developed a fairly efficient rhythm, understanding the scope of each project and allocating resources accordingly. However, as our studio grew in terms of clientele and the complexity of projects, we started to face unforeseen challenges. Changes in creative workflows, unexpected revisions from clients, and the unpredictable nature of creative work started to affect our carefully planned timelines.

The problem became clear when an underestimation of a major project’s scope at the beginning of the year led to a big overrun in both time and budget. The effects were felt across the studio, straining our resources and affecting our reputation for timely delivery.

This prompted a search for a more accurate, data-driven approach to project management. The idea was to move from reacting to problems to a proactive, predictive model that could provide more accurate, real-time insight into project timelines and potential issues. The objective was clear: to use technology to refine our project estimation and management processes, ensuring a smoother workflow that could handle the dynamic and creative nature of our studio while meeting the punctuality expectations of our clients.

It was around this time that the idea of Artificial Intelligence (AI) as a solution caught our attention with the release of ChatGPT. The thought of using AI to analyze past project data, predict timelines, and provide useful insights was intriguing. However, the workflow was tedious, to say the least.

The introduction of Tana AI was a sign of hope, offering a solution that seemed perfect for our unique challenges. Eagerly, I tried to integrate a series of commands into our workflow, with the vision of creating a strong, AI-powered project management system. The initial attempts were promising but fell short of our expectations in some areas.

But the real change came with the advent of GPT-4, which, when combined with Tana, provided a level of predictive accuracy and insight that was revolutionary.

Unveiling the Node Filter Mechanics

In the initial node filter setup, we instruct the + Prediction Button to appear and become accessible solely when there’s a precise match between Prediction field / Not Set and Task Weight / Set.

Upon a successful match, the button materializes. When clicked, it attempts to execute the foremost command in the queue, which is UI/UX Prediction.

Now onto the second filter. Here, we are instructing the + Prediction Button to execute UI/UX Prediction only if the Type of Task field matches. If there isn’t a match, it will bypass this and move on to the next available command in the list, which in this scenario is Logo Design Prediction, continuing this process until it locates a match.

Now, you might notice that there are three more commands nested under UI/UX Prediction, and each of these commands has an additional Node filter tucked under it 😅.

Context: I utilize the Task Weight field on work todos to further categorize tasks post-completion. In a nutshell, if a task took us 45 days to complete, it’s undoubtedly a heavyweight task and will receive a high score of 99 👹. This necessitates three additional commands to accommodate all our scenarios.

By nesting a Node filter under each of our Task Weight options, we are instructing the + Prediction Button to execute only when all three filters find a match.

We replicate this process for each Task Type, eventually crafting our very own 4-headed command 🧌. It diligently searches through the 12 options for that magic trio of matches, and upon finding them, it triggers the AskAI Prompt to run. 🪄

So what?

The impact on our studio has been huge. The predictive power of a single command has enforced our scheduling, instilling a newfound confidence in committing to deadlines. With a more predictable workflow, we’ve ventured into new client territories, no longer merely hoping, but knowing we have the time to deliver.

This evolution in our project management approach, powered by Tana and GPT4, is more than a technical upgrade—it’s a leap toward fulfilling our studio’s creative aspirations while meeting the pragmatic demands of project timelines.

My exploration in AI-driven project management continues to evolve every day. Each iteration not only sharpens our scheduling foresight but opens doors to broader possibilities—like automating client interactions or leveraging other AI features to refine our workflow.

The path forward is thrilling. With every new insight, we’re not just meeting deadlines but broadening our scope, crawling closer to a future where AI becomes a cornerstone in delivering exceptional design, on time, every time, while continually uncovering new avenues to enhance our studio’s operational rhythm.

The journey to creating commands like this was fueled by the folks at the Tana Nodes Discord Server. Their questions, challenges and insights pushed me to explore more and more everyday. I owe them my monumental gratitude.

If you want to dive deeper into the world of Tana Commands and AI, check out my extensive collection of resources. Whether you’re just starting out or looking to enhance your existing knowledge, there’s a wealth of tutorials, templates, events and community discussions awaiting you.