90% Customer Retention — Mill Pond Research's story with TypingMind

It's great to hear TypingMind fits so well with Ask Alpha! Many of our users are also taking advantage of multiple AI models, built-in plugins, prompt libraries, and AI Agent collections for their team members.

Could you elaborate on how TypingMind is integrated into your daily operations?

We offer a managed premium business-focused AI chat experience for professionals. Through our training data and prompting methodologies, our users are more confident in the accuracy and quality of our AI outputs. Many of our users come from other AI systems seeking out a better AI system, and Ask Alpha delivers. Our users range in professions – copywriters, lawyers, students, entertainment professionals, creatives, government contractors, and more. They use it as their digital multi-tool – for asking questions, searching the web, document creation, image generation, code generation, brainstorming, data analysis, creative assistance, translation, reformatting documentation (from outline to presentation), summarizing documents, creating knowledge bases, and more.

You seem to have a solid strategy for implementing TypingMind for your users/clients. Your approach of combining TypingMind's flexibility with custom training data and tailored methodologies is exactly what we encourage our users to strive for.

Among all the features available within TypingMind, which one has proven to be the most valuable in improving Ask Alpha's capabilities?

Many of the features in TypingMind are unrivaled in the AI space. We believe the ability to switch between AI models is crucial. We leverage the custom prompt library constantly – updating prompts and tweaking them as best practices get updated. Our AI Agent library is now at over 50 agents, each with custom instructions, knowledge, and dynamic information pulled in through APIs. We love the responsive UI – allowing for seamless use on desktop or mobile. The new Prompt Chaining is an exciting and very powerful new feature - we're still exploring how to fully utilize its functionality.

Ask Alpha implements Prompt Chaining on TypingMind

As AI models continue to evolve, the ability to adapt and switch between them is crucial for delivering optimal results to your clients.

And wow, the number of your AI Agents is impressive! Could you share more details about some of these AI agents and the roles they play for your team members?

Certainly! We've developed several AI agents tailored to meet our needs.

For example, our "US News Reporter" agent provides incisive and insightful news, helping clients stay informed with the latest developments. The "HTML Wizard" is designed to assist in creating immersive HTML content quickly and efficiently. Meanwhile, our "Professional Writer" agent supports content creation and copywriting, offering high-quality writing assistance for a range of professional purposes.

Ask Alpha AI Agent Library

Would you say TypingMind has fully met your needs so far?

Yes! We love TypingMind and think it's the ideal system to build a custom AI workbench on.

Thank you, Pete! How has TypingMind contributed to your overall business goals? Have you seen any measurable impacts?

Our customer retention rate is higher than 90%. We attribute this to excellent software, provided by TypingMind, as well as our "Hero-Making" client success services.

Ask Alpha Analytics Dashboard on TypingMind

I'm curious about your decision-making process when choosing TypingMind. Did you evaluate other solutions before settling on TypingMind? What made it stand out from other options in the market?

Yes, and we're constantly evaluating other systems. We continue to use TypingMind for its versatility, flexibility, and customization options. It feels like a "Wordpress" for Chat AI.

The comparison to WordPress is interesting - it really highlights the flexibility and customization potential of TypingMind.

As you continue to push the boundaries of what's possible with the platform, have you encountered any limitations or areas where you'd like to see improvements? Are there any features you're hoping to see in future updates?

Yes, the ability to concurrently prompt multiple models/agents and bring the results back into a main thread. See Beam: Better Decisions, Lower Risk, with Multi-Model AI Reasoning – big-AGI. I think with TypingMind, this feature could become even more powerful leveraging Agents versus just models. (Imagine being able to call in multiple agents into a single thread or working with one agent and pulling in a second one). This would be a killer feature.

• Allow for "Custom Model Settings". For example, GPT-4 now has a maximum token output of 16,000+. When I switch between GPT-4 to Claude 3.5 Sonnet without adjusting the model settings, I'm met with an error message saying the token limit is too high. I should be able to save settings to specific models versus just the "Default Settings".
• I'd love to be able to customize the default AI agents provided by TypingMind.
• I really want a "BrightData" plugin which allows the user to add any of the BrightData APIs for contextualizing the prompt with information from the web. This would be huge!

Given your positive experience and the success you've had with TypingMind, would you recommend it to other businesses looking to implement AI solutions? And if so, what advice would you give them?

Absolutely, we feel that we're not customers but partners of TypingMind. They focus on the software, and we focus on optimizing the customer/user experience to be the best AI chat experiences on the web. Many users find the process of setting up and managing an AI system to be complex – this is where Mill Pond Research is an excellent partner in building intelligence systems for organizations.

Thank you so much for sharing valuable insights, Pete!