Skip to main content

Ongoing Maintenance and Development

Ongoing Maintenance and Development of Your AI Expert: A Never-Ending Process of Improvement​

Your AI expert is not a product you create once and then forget about. To maintain its value, relevance, and competitiveness in the rapidly evolving world of artificial intelligence, it requires a certain level of ongoing care and strategic development on your part. This process is key to long-term success and client satisfaction.

Regular Updating and Supplementing of Your Know-How (keep the AI "fresh")​

Your expertise and knowledge are constantly evolving. You learn new things, create new content, gain valuable practical experience, and respond to current trends in your field. For your AI expert to remain relevant and provide maximum value, it's essential to regularly "pass on" these new insights to it.

  • Why is this important:

    • Maintains the currency and accuracy of information the AI provides.
    • Increases the value of the AI expert for your clients, as it reflects the latest insights.
    • Allows the AI to better respond to new types of queries and situations.

  • How often to update:

    • The frequency depends on the dynamics of your field and how often you produce new content.
    • The recommended interval can be once a month, quarterly, or whenever you create a significant new block of know-how (e.g., a new online course, e-book, series of webinars).

  • What specifically to do:

    1. Gather new materials: Identify your latest articles, webinar/podcast transcripts, updated workflows, new case studies, answers to frequently asked questions, etc.

    2. Upload them to Google Drive: Place these new files in the appropriate subfolders in your main BuddyPro folder on Google Drive, specifically in the sources folder (e.g., in sources/Files, sources/Texts, or in the sources/URLs/URL SOURCES file).

    3. Review and delete outdated materials: Go through the content in the sources folder and remove or update files that are no longer valid, accurate, or relevant.

    4. Run synchronization: After any changes to content in the sources folder (adding, deleting, editing files), it is essential to go to the Dashboard to the Know-how section and perform a know-how update.

      This ensures that all changes are reflected in your AI's knowledge base. Without running it, the AI wouldn't know about new or changed materials.

Ongoing Tuning of System Prompt and Onboarding (responding to feedback and new insights)​

The personality, behavior, and first impression of your AI expert are not set in stone. Based on user feedback, your own interactions with the AI, or changes in your marketing strategy, it may be necessary to fine-tune these key aspects from time to time.

  • Why is this important:

    • Allows you to respond to user needs and preferences.
    • Helps correct unwanted AI behavior that may appear.
    • Ensures that AI communication aligns with your current brand and goals.

  • Tuning the system prompt:

    • When: If you find that the AI is not behaving consistently, using an inappropriate tone, or you need to adjust its basic instructions.
    • What to adjust: In the Dashboard in the System Prompt section, you can:
      • Refine the AI personality definition (e.g., friendlier, more formal, more empathetic).
      • Add new behavior rules or modify existing ones.
      • Improve or supplement disclaimers (e.g., for legal, financial, or health topics).
      • Add a "Critical errors and their solutions" section to correct specific, recurring errors in AI behavior.

  • Tuning onboarding messages:

    • When: If you want to improve the first impression for new users, increase their engagement, or update introductory information.
    • What to adjust: In the Dashboard in the Messages section, you can:
      • Rework welcome message texts to make them more attractive, clearer, or to include current information (e.g., link to a new onboarding video).
      • Adjust the timing or content of automatic onboarding messages that are displayed to the user after a certain number of interactions.

Considering Regeneration of Expert Sub-Roles (after very significant changes in know-how)​

Expert sub-roles are specializations of your AI that are automatically generated based on uploaded know-how. They allow the AI to better respond to specific queries within your expertise.

  • When is role regeneration appropriate:

    • If you make a truly major and extensive update or restructuring of your know-how. This may include:
      • Adding content from a complete new flagship course covering many new topics.
      • Significant revision of your core methodology or key concepts.
      • Adding a large amount of new content that fundamentally changes the breadth or depth of your AI's knowledge.
    • In such cases, the originally generated roles may not optimally reflect the current state of your knowledge. Regeneration will help them better "form" according to the new comprehensive know-how.

  • How to perform complete role regeneration:

    1. First, make sure all new know-how is uploaded in the sources folder on Google Drive.

    2. Run in the Dashboard in the Know-how section, a know-how update and wait for its successful completion.

    3. Then go to Google Drive to the ROLES folder of your BuddyPro instance.

    4. Delete all Google Document files that contain definitions of individual roles (typically they have names like Role_RoleName.gdoc).

    5. IMPORTANT NOTE: In the ROLES folder, DO NOT DELETE the list_of_roles.gdoc file! This file contains only a list of proposed role names and is needed for the regeneration process.

    6. After deleting the role definition files, run in the Dashboard in the Roles section, a role update.

      This process may take longer depending on the amount of your know-how. The AI will go through the list of role names and create a new, current definition for each based on all available know-how.

  • When role regeneration is NOT usually necessary:

    • For smaller, ongoing additions to know-how (e.g., adding one article, updating a few paragraphs, a small video).
    • In these cases, the AI will have access to new information even without regenerating roles; it will use them when searching for relevant snippets of know-how for an answer. Roles primarily help the AI specialize its behavior.

Tracking News and Improvements to the BuddyPro Platform (stay informed!)​

The BuddyPro platform is not a static product. Our development team is constantly working on improving it, adding new features, optimizing performance, integrating the latest AI models, and fixing any bugs. To fully utilize the potential of your AI expert, it's important to stay informed.

  • Why is this important:

    • New features can bring improvements for your AI expert or make its management easier.
    • Optimizations can improve speed, response quality, or reduce costs.
    • You'll be informed about any changes that could affect the functioning of your instance.

  • How to stay informed:

    • Telegram group BuddyPro Owners: This is your main information channel.
      • Regularly follow the News and announcements channel. Here we will communicate all important platform updates, new features, planned maintenance (if exceptionally necessary), and other relevant messages.
      • Engage in discussions in other group channels where you can share experiences and get tips from other owners.
    • Webinars, masterclasses, Q&A: Occasionally, we may organize online meetings for BuddyPro instance owners, where we will present news, share strategies, or answer your questions. You will be informed about these events.

  • Using new features:

    • When we introduce a new feature or improvement, think about how it could enrich your AI expert, simplify your work, or improve the user experience for your clients.
    • Consider implementing new capabilities to keep your AI expert at the cutting edge.