Let’s avoid this.
In the last Part we built an early version of our tone of voice tool - using Projects.
We kept it nice and simple using Projects so we could i) create a knowledge base and ii) create an instruction prompt off the back of the knowledge base. Solid.
If you are using your tone of voice tool personally - just for you - then you can actually skip to Part 5. You’ve got what you need!
If however you want to expand who can access the tool (say, your marketing team) OR if you are building for an organisation then this Part is vital. We’re talking deployment.
The perfect tone of voice model isn't worth much if the right people can't access it when they need it.
Let's get started:
Before we dive into specific platforms, we need to answer some crucial questions about how your tone of voice system will be used. The right deployment option depends entirely on your specific needs. No one size fits all here.
Let's start with a decision framework to guide your choice:
To make this process easier, I've created (you guessed it!) a prompt that will give you personalised deployment recommendations specifically for our voice system approach:
You are an AI deployment strategist specialising in brand voice systems. Help me determine the optimal deployment option for my AI brand voice model based on my specific needs.
Context: I have built a tone of voice replication system consisting of:
1. A knowledge base of multiple uploaded documents (potentially hundreds of files)
2. Custom instructions derived from analysing these documents
3. A system that references specific examples from the knowledge base when generating content
Ask me the following questions one by one, waiting for my response to each:
1. Who will be using this voice system? (Options: Just me personally, My business team, A client's team, Other)
2. How many people need access to this system? (Options: Just 1, 2-5 people, 6+ people)
3. What is your budget for this deployment? (Options: Minimal/free options only, Moderate budget, Enterprise budget)
4. What is your technical comfort level? (Options: Non-technical, Some technical knowledge, Technical expert)
5. How will the content be used? (Options: Internal use only, Public-facing content, Both)
6. How many documents does your voice knowledge base contain, and what's their approximate total size? (Options: Few small documents, 10-50 medium-sized files, Large document library)
7. How important is the ability to easily update your voice knowledge base over time? (Options: Not important, Somewhat important, Very important)
Based on my answers, recommend the best deployment option(s) for my specific voice system needs. For each recommendation, include:
- Platform name
- Why it's suitable for my voice system implementation
- How it handles my knowledge base and custom instructions
- Basic setup steps
- Limitations regarding document handling and reference capabilities
- Approximate cost
Present the top 3 options in order of recommendation, with clear reasoning for each ranking.This prompt will help you narrow down your options based on your specific needs. Use it with any AI assistant (it doesn't need to be your voice model!) to get personalized recommendations.
Now let's look at the main deployment options available. Obviously there are more than these (and the prompt above will help you discover them) but let’s orientate ourselves to the main options.
Best for: Personal use, individual creators, simple deployment needs
Pros:
Cons:
Setup process:
Cost: Free (with limited capabilities) or $20/month (ChatGPT Plus)
Best for: Small teams, collaborative workflows, long-form content
Pros:
Cons:
Best for: Quick deployment, non-technical users, client deliverables
Pros:
Cons:
Popular options:
Cost: Typically $20-500/month depending on features and scale
Best for: Technical users, complex needs, integration requirements
Pros:
Cons:
This is all in flux obviously. So what works now as a deployment option may not be ideal 6 months from now.
Generally though with AI everything is becoming easier, more accessible and better. Change is good!
As you decide on your deployment option, keep in mind that this space is evolving rapidly:
All you can do right now is build using whatever makes sense at this point in time.
Thankfully your knowledge base and custom instructions will remain useful on different platforms in the future - all that prep work we completed in the last Parts remains valuable.
In our final Part we'll focus on training and refinement. How to make your tone of voice model better by giving it feedback.
You'll learn how to systematically improve your tone of voice system through feedback, testing, and iteration. For either your personal tool or for clients.