I’ve been responsible for Trivera’s email marketing since day one. While Trivera’s talented team of superstars are the ones who do all the work to help our clients succeed. I’ve hung onto the duties of communicating with our audience and shaping how we show up in their inbox. List management, monthly newsletters, and now our Deep Dive episode alerts. For years, it was a manual process. Every month, start fresh, build it, send it.
When we trained Webster, our AI assistant, I started using it to help build those newsletters. Same structure every month. Same layout. Same need for clean HTML. Different content.
At first, it worked great. I kept everything in a single chat thread. It felt efficient.
Until it didn’t.
The thread got longer. The outputs started drifting. And eventually, I hit the wall. Maximum chat length. Time to start over.
That’s when it clicked. The problem wasn’t the AI. It was how I was using it.
That’s when I discovered projects.
Why This Matters More Than Prompts
We’ve talked before about how to write a strong AI prompt. And those matter. A good prompt at the start of a chat can produce great results.
But here’s the problem.
Prompts don’t scale inside long conversations.
As the chat grows, that original clarity gets buried. Context drifts. Instructions get diluted. Outputs get inconsistent.
So you compensate.
You rewrite the prompt. You restate the rules. You try to steer it back.
And over time, that becomes the work.
Projects solve that.
They take what worked at the beginning of a chat and turn it into a structured, repeatable environment where that clarity doesn’t get lost.
Different AI platforms use different names for essentially the same idea. In ChatGPT and Claude, they’re called Projects. In Perplexity, they’re called Spaces. The label changes, but the goal is similar: create a structured workspace where recurring work, source material, and instructions can live together instead of getting buried in a long chat.
How We’re Using Projects at Trivera
At Trivera, this isn’t experimental. It’s operational. What started out as a fix for one problem is now a system we use across the business.
We’re now using project-based workflows for:
- Blog drafting and publishing preparation
- Podcast production and AI agent creation (for us and for clients)
- Structured content generation across campaigns
- Recurring marketing tasks and reporting
- Sales presentations and project process documentation
- Contracts and business correspondence
The goal isn’t just to produce content faster.
It’s to reduce the friction around how marketing work actually gets done.
And one of the advantages of how we’ve implemented this with Webster is that these projects aren’t isolated. We can share them, adapt them, and reuse them across the team.
So whether someone is working on an email, a campaign, a report or a presentation, they’re not starting from scratch. They’re starting from something that’s already been built, tested, and refined.
A Practical Step-by-Step: How to Build a Project
At Trivera, projects have become a core part of how we get work done. But we don’t work in a vacuum. Our clients have talented marketing teams that we collaborate with every day, and they need to be able to use AI just as effectively on their side of the partnership.
This is where projects become especially valuable. Not as a theory, but as a practical way to improve how real marketing work gets done.
If you want to start using projects effectively, don’t overcomplicate it. Start with something you already do on a recurring basis.
Step 1: Pick the Right Use Case
Look for work that is:
- Recurring (monthly, weekly, ongoing)
- Structured (follows a format or pattern)
- Context-heavy (requires consistency in tone, style, or output)
If you’re repeating yourself in chats, that’s a good candidate.
Step 2: Create and Configure the Project
In ChatGPT, this part is simple but important.
- Name the project clearly (e.g., “Monthly Newsletter Builder”)
- Decide whether the AI should rely only on your uploaded sources or also use web knowledge
- Select any available domain or usage settings if relevant
This isn’t where the magic happens. It just sets the container.
Step 3: Write Better Instructions
This is where most of the value comes from.
Start with the best prompt you’ve already used. Then refine it.
Good project instructions should define:
- The purpose of the project (what it is supposed to produce)
- The audience (who the output is for)
- The tone and voice (formal, conversational, technical, etc.)
- The structure of the output (sections, formatting, order)
- Specific requirements (HTML, word count, formatting rules)
- What “good” looks like (clarity, accuracy, usability)
- What to avoid (fluff, repetition, unsupported claims)
If your outputs are inconsistent, the issue is usually here.
Step 4: Add Source Material
This is what grounds the project and improves consistency.
- Good sources might include:
- Previous newsletters or blog posts
- Brand guidelines or tone-of-voice documents
- Sample outputs that represent “done right”
- Product or service descriptions
- Campaign messaging or positioning documents
The goal is simple: give the AI something to anchor to so it doesn’t have to guess.
Step 5: Test and Refine
Run a real task through the project.
Then adjust:
- Tighten instructions where outputs are vague
- Add sources where context is missing
- Remove ambiguity wherever possible
This is not a one-and-done setup. It gets better with use.
The Cheat Code: Don’t Start From Scratch
You probably already have what you need.
Go back to a chat you’ve been using for something important. Especially one that’s gotten long, messy, or inconsistent.
Then ask the AI to help you convert it.
“Take this chat and help me turn it into a project. Define the purpose, create clear instructions, and suggest source material.”
You can also ask it to:
- Summarize the original goal of the conversation
- Extract the best version of your prompt
- Identify patterns in what you’ve been asking for
- Recommend what should become instructions vs. sources
You’re not starting over. You’re upgrading how the work is structured.
What This Means for You
If your team is still working primarily in chat threads, you’re probably experiencing more friction than you need to.
Look at your recurring work first. The things you do every week, every month, every campaign cycle. That’s where projects deliver the most value.
You don’t need a better AI.
You need a better way to use it.
At Trivera, we’ve always believed in leading with strategy, not chasing tactics. AI is no different. It’s easy to experiment, to dabble, to try a few prompts and see what happens. Most teams are doing that right now.
But the real advantage doesn’t come from experimenting. It comes from organizing.
It comes from taking the work you already do, defining it clearly, structuring it intentionally, and building systems around it so it can be executed consistently.
That’s what projects allow you to do.
Not just use AI, but integrate it into how your marketing actually operates.
Because the teams that win with AI won’t be the ones chasing the latest feature.
They’ll be the ones who apply it strategically to the work they already need to get done.
Where Trivera Fits In
For Trivera RSA clients, this is an area where we can help teams improve how they work with AI.
Not just using it, but structuring it around real marketing workflows.
And if you’re not there yet, this is part of what modern agency partnership should look like.
