How To Deploy AI Across Your Company 🤖

I know, yet another AI post? We’re all sick of hearing it or about it, right? And, truthfully I’m really not the biggest AI fan myself! However, If you’ve tried to expand AI adoption across your organisation, you've probably encountered a familiar pattern: inconsistent results, frustrated team members, and endless prompt refinements.

This challenge multiplies across your organisation as different team members develop unique approaches to prompting and AI projects, leading to inconsistent results and inefficient workflows. Trust me, I know about this first hand!

So let’s jump into the AI solution that is a systematic prompt framework to scale from a single source of truth.

The AI Strategy Canvas For Scalable AI Prompt Engineering 👷🏻

The AI Strategy Canvas is built on a two-part foundation of prompt formatting and dynamic variables that can be quickly dropped into a prompt. This is easily achieved by using delimiters to open and close a prompt block. 

Imagine you want to write an email using Claude. Start your block with all caps and a colon.

Then, enter the name of a target audience persona you've already fully defined and stored in a library where your AI can retrieve information. This is your variable.

Using dynamic variables that pull standardized details from a library ensures that everyone can easily include the correct information in their requests because there is a single source of truth. You can store these variables in Notion or similar tools your AI can connect to, allowing quick substitution (AKA hot-swapping) for different use cases.

Close your block with a forward slash.

For example, this prompt tells Claude that you want it to generate content for prospects who fall into your Target Audience A–without you having to copy and paste lengthy persona details.

AUDIENCE:
The audience is: ${target_audience_a}
/AUDIENCE

The 9 Blocks For Scalable AI Prompts 🧱

Each block will have multiple variables you can swap in for specific uses. Give each varia a name–paying special attention to taxonomy, and store them in your prompt library for hot-swapping. For example:

• target_audience_a

• target_audience_b

• target_audience_c

Building Block 1: Target Audience Definition: Using psychographics, demographics, economics, and sociographics, build a complete profile for each target audience.

Building Block 2: Company Context Integration: Create multiple versions of your company context, from a one-paragraph summary to comprehensive documentation that can be used at various customer journey stages.

Building Block 3: Products and Services: Maintain two distinct versions of your product information to detail the value you deliver to customers: a concise promotional description for basic tasks and a comprehensive document detailing features, benefits, and technical specifications.

Building Block 4: Dynamic Context Integration: Think of this as a situation-specific catch-all block. For example, you might need the AI to reference:

• Key points discussed in a review meeting

• Specific challenges your team identified

• Customer feedback

• Current business objectives

Building Block 5: Role Definition: Be very clear about the roles your team will need and how you want the AI to behave within the scope of each role. For example:

• You are a data analyst specialising in customer retention.

Building Block 6: Style and Voice Calibration: Create several style and voice blocks so that you're using the best fit for every need, such as internal emails, prospect emails, and customer emails.

You can develop specific parameters to control everything from tone to sentence length in AI outputs. For example:

• Humor Level: 4 on a scale of no humor (0) to extreme humor (10)

You can also maintain an exclusions rule that gives the AI a list of words and phrases your company doesn't want to see in their content.

Building Block 7: Resource Integration: This block connects your AI tools with the information and tools they need to complete your request. These resources might include:

• Meeting transcripts

• Reference documents

• External URLs

• API data feeds

Building Block 8: Rule Setting: Beyond the basic guidelines from Block 6, these are detailed rule sets that establish guardrails for data security and copyright infringement.

Building Block 9: Request Formulation: Your actual request for output should build upon all previous blocks to create a clear, actionable instruction for your AI tool–what you want the final output to look like. For example, if you want a fully fleshed-out article, use the word ‘Write" instead of “Outline.”

So, I hope this all will help you get AI deployed into your business in a sufficient manner.

Q&A

• So, are you ready to harness the power of using AI in your business?

• Do you have any pain points I can help you with smoothing out?

• What do you think you could utilise AI for and how would it impact your business?

Whenever you’re ready, here’s how I can help you:

1. Book a {10 MINUTE} Business Power Call with me, to arrange working together to make your social media videos stand out online as I tweak your approach, and analyse how to improve your current video, business strategy or marketing.

2. Don’t have time to create great videos yourself? Focus on your business, whilst we focus on creating the videos that you’ve always dreamed of. Get in touch to find out more.

Until next Friday, love to you all!

My Partner Services
Sky Adsmart  | Cheekymonkey Marketing  |

Fire an email back to me at [email protected] and let’s keep the chat going 👌🏻