Traffic Builders leads the way in AI-driven marketing with TypingMind

We try to incorporate supporting our consultants with their specific tasks into TypingMind Custom by creating custom plugins to help with some of our more time-consuming tasks so TypingMind becomes more of a one-stop-shop for us.

We use the app for various purposes, but personally, as a Content Marketer and Copywriter (and one of the AI team members), I use it for seeking inspiration when writing texts, proofreading my content, and brainstorming ideas related to content marketing. Ultimately, our use of AI in content creation is focused on intelligently harnessing its power to make our content even better, more relevant, and more valuable to the target audience.

By leveraging AI, we aim to accelerate and optimize our processes, automating peripheral tasks that would otherwise consume a significant amount of time. This allows us to dedicate more time to creativity and proactivity. The key for us is to improve our (working) lives and work (for clients) with the use of AI. TypingMind is supporting us in achieving this by providing access to all the best AI models for our employees.

As a team, we recognize the importance of developing and maintaining the knowledge and skills necessary to work effectively with AI. This involves experimenting, learning from best practices, and collaborating closely with the AI experts within Traffic Builders.

We also strive to find the right balance between AI and human creativity and empathy, ensuring that AI serves as a tool to enhance our results, rather than using AI for its own sake.

Yes, TypingMind has fulfilled our needs at Traffic Builders. It is a pleasant platform to work with and has given us the opportunity to quickly test various new language models and AI plugins. We have also developed several plugins ourselves, such as a PageReader, CookieChecker, and a PageSpeed tool.

Moreover, TypingMind has helped us further implement AI solutions for our clients. We offer basic AI training sessions to our clients and provide more in-depth collaboration sessions specifically focused on setting up AI within a company, creating AI Agents, and more.

We are currently working on AI cases that we hope will be award-winning, as the results have been very promising.

Multiple LLMs and plugins enhance AI solutions for Traffic Builders' clients

We've been developing several AI Agents that assist with various tasks, such as Researcher, AdCopy, Account Manager, Advertiser, Designer, Sales Manager, and Copywriter.

For example, the Art Director helps conceptualize and develop creative ideas following the REAN method and the Copywriter can write and paste texts based on briefs from the Art Director.

AI agents have helped with various tasks for Traffic Builders' members

One interesting use case we have implemented in TypingMind involves creating an AI Agent modeled after Steve Jobs.

Our digital SEO consultant, Robbert van Ekris, shared this innovative approach:

🍏 Steve Jobs in UnboundGPT

Steve Jobs was known for his magical showmanship. My attempts at presenting are more of a tragedy than magic 🎭.

So I thought: "Why not invite Steve Jobs for a guest appearance?"

I created an AI Agent in UnboundGPT, linked it to some internal knowledge about us and our presentation, and cloned Steve Jobs' voice using ElevenLabs.

Test it yourself by inviting him (the AI Agent is shared with everyone) for a chat about Tije and Robbert:

• Let's briefly introduce Tije Vlam and Robbert van Ekris.
• Please provide a two-sentence summary of the first part of the presentation done by Robbert and the second one by Tije at Overstappen.nl.

During my presentation, I asked these same questions live.

This AI Agent demonstrates how we leverage TypingMind to create engaging presentations and enhance our communication strategies within Traffic Builders.

Steve Jobs in UnboundGPT

We believe that each used token represents a form of automation or improvement, which can be seen in our usage statistics. The more tokens we use, the greater the efficiency gains we achieve through automation and enhanced processes.

One major advantage is that with TypingMind, there was zero implementation time required to get the AI team on board with UnboundGPT. When a new model becomes available, integrating it takes virtually no time at all. This has greatly increased our innovation capacity and accelerated development times. If we had attempted to build something similar ourselves, we would still be far from completion.

Additionally, not being tied exclusively to OpenAI gives us greater flexibility and prepares us better for future advancements.

Traffic Builders leverages multiple LLMs in Typingmind for chatting

Some of our colleagues were already using multiple paid AI models, such as a GPT+ subscription, both from their study budget and out of personal interest. However, TypingMind is the first solution we have tested that combines the different models in such a clear, organized, and - most importantly - effective way. This is much faster than testing various separate tools and models independently, as TypingMind makes it easier to switch between models and settings.

We’ve also looked into several open-source ChatGPT clones. But none of them worked as well (at the time) or were model-independent.

The primary goal behind having our own AI chat interface is ensuring unlimited access to state-of-the-art models like GPT-4o Turbo, Claude Opus, and Gemini 1.5 Pro via our API keys.

It would be beneficial to focus on quickly integrating new models and adding features that make it easier to edit and interpret our own data from sources like Google Drive and Microsoft OneDrive.

Make the “Temperature” slider clearly accessible during a chat, and allow the possibility to change advanced settings for AI agents, including at least Temperature and Top_p.

Additionally, as a feature request, we would like the ability to directly engage specific assistants (from OpenAI), GPTs, or data interpreters.