Creating an AI Assistant: How to Get Started

Wondering if AI could help you make better decisions, faster? Thinking of using AI to help you perform better? 

In this article, we explore how to create an AI assistant.

Understanding AI Assistants

An artificial intelligence (AI) assistant is essentially a digital helper that can take on various tasks for you. It's like having a virtual employee to whom you can delegate work. While AI assistants are simpler than human assistants in some ways, they're actually smarter in others. For example, an AI assistant has basic knowledge about almost every topic imaginable, from history and science to current events and pop culture. This broad knowledge base is something most human assistants wouldn't have. 

However, it's important to understand that AI assistants are not meant to replace human workers completely. Instead, they're designed to complement and enhance human capabilities. By taking care of time-consuming, repetitive tasks, AI assistants free up humans to focus on higher-level, creative work that requires critical thinking and emotional intelligence.

For instance, an AI assistant could easily summarize a long article into bullet points, saving you the time of reading through the entire piece and extracting the key information yourself. It could also quickly generate ideas for blog post titles, social media updates, or email subject lines based on a given topic—something that might take a human much longer to ideate. 

At the same time, an AI assistant couldn't replace the strategic thinking required to develop a full content marketing plan or the interpersonal skills needed to build relationships with customers and colleagues. Its outputs would still need to be reviewed and refined by a human to ensure they align with brand guidelines, target audience preferences, and overall business objectives.

In this way, AI assistants are best considered powerful tools that can streamline and scale certain aspects of knowledge work. However, they still require human oversight and collaboration to be used effectively.

The Time to Start Is Now

If you're not already using AI in your work, it's easy to feel overwhelmed or intimidated by the sheer pace and scale of change in this space. Every week, it seems, a new breakthrough or product launch promises to be a game-changer. How can you possibly keep up?

The key is to start small and start now. You don't need a PhD in computer science or a huge budget to begin experimenting with AI in your job. Thanks to user-friendly tools like ChatGPT, DALL-E, and Midjourney, anyone can start playing with powerful AI models and seeing what they're capable of—often for free!

The best way to get comfortable with these tools is to dive in and use them. Pick a simple task or process in your work that you'd like to streamline, and challenge yourself to find the right AI platform or tool to help. It could be something as basic as using ChatGPT to help you brainstorm ideas for your next blog post or Midjourney to generate visual inspiration for a design project.

As you experiment with different prompts and settings, pay attention to what works and what doesn't.

#1: How to Prompt Your AI Assistants

The key to getting started with AI is to experiment with simple prompts. 

A prompt is essentially an instruction or question you give to the AI technology to get it to perform tasks. Igor Pogany, founder of AI Advantage, suggests beginning with a straightforward prompt like “What is the meaning of life?” or “Generate an essay about penguins.” The idea is to start with a basic command and see what output the AI generates.

Once you have a general sense of how the AI responds, you can start making your prompts more specific and tailored to your needs. For instance, instead of asking the AI to summarize an article, you might say, “Summarize this article in five bullet points, written in the style of an 11-year-old.” This additional context helps the AI understand what task to perform and how you want it done.

Here's another example: Instead of just saying, “Write a blog post about the benefits of meditation,” you might prompt the AI with something like, “Write a 1,000-word blog post discussing the top 5 scientifically proven benefits of meditation for stress relief. Use short paragraphs, a friendly conversational tone, and include practical tips a beginner could use to start meditating today.”

The more specific your prompt, the more targeted and useful the AI's output will likely be. You could even include examples of the format or style you want to help the AI better match your expectations. 

The goal is to find prompts that generate results you can actually use in your work, with minimal editing required. This might take some trial and error, but it's a crucial step in making AI a valuable tool in your toolkit.

Whenever you discover a prompt that works well, save it somewhere so you can reuse and build upon it easily. Igor recommends keeping a spreadsheet or document with your favorite prompt templates, categorized by topic or use case. 

Over time, as you collect more and more effective prompts, you can start automating entire workflows and creating more advanced AI assistants tailored to your specific needs. For example, you could create a prompt template for generating social media updates whenever you publish a new blog post or for drafting responses to common customer service inquiries.

#2: How to Give Your Own AI Assistant Feedback to Improve the Output

As you experiment with different prompts, pay close attention to the quality of the AI's outputs. Remember, just because an AI can generate human-like text quickly doesn't mean that text is always accurate, relevant, or aligned with your goals. It's ultimately up to you to review the AI's work with a critical eye and decide if it meets your standards.

When the AI's output misses the mark, don't hesitate to give it direct, constructive feedback. Igor suggests using a numeric scale, like “I'd rate this draft a 5 out of 10,” to clearly communicate your assessment. Then, explain exactly what the AI got wrong or left out so it can learn to do better next time.

For example, imagine you asked the AI to write a product description for a new sneaker your company is launching. If the output doesn't mention key selling points like the shoe's lightweight design or eco-friendly materials, you might say something like:

“This description gets a 6/10. You covered the basics of the shoe's style and color options but didn't highlight the unique value propositions that really set it apart. Please try again, and this time make sure to mention:

• The shoe is 20% lighter than the leading competitor 
• It's made from 100% recycled plastic bottles
• For every purchase, we donate $10 to environmental causes”

The more specific and actionable your feedback, the faster the AI will learn your preferences and start producing better results with each iteration. Of course, you can also praise the AI when it does a great job of reinforcing what success looks like.

Remember, the goal isn't to make the AI feel bad—it's an algorithm, not a person with feelings to hurt. The goal is to train the AI to be a more useful tool for your needs. Think of it like coaching a new employee—your direct guidance is essential for getting them up to speed.

#3: How to Ask Your Own AI Assistant the Right Questions to Improve the Output

In addition to giving feedback on the AI's outputs, another key skill in working with AI assistants is learning to ask them the right input questions. The clearer and more detailed your prompts are, the better the AI can give you the results you're looking for.

One powerful technique Igor recommends is asking the AI to prompt you for the information needed to do a task well. So instead of just saying, “Please write a press release for our new product,” you might say something like:

“I need you to write a press release announcing the launch of our new product. To do this effectively, please ask me any additional questions you have about the product, the launch plan, our target audience, or our overall goals. I'm happy to provide more context to help you create the best possible press release.”

By inviting the AI to ask clarifying questions like this upfront, you're transferring your domain knowledge to the AI, helping it understand the specific requirements and nuances of the task. The AI might follow up with questions like:

• What are the key features and benefits of this new product?
• How is it different from or better than competing products on the market?
• When and where will the product be available for purchase?  
• Who is the ideal customer for this product?
• What key message or emotion do you want this press release to convey?

Armed with your answers to these questions, the AI can generate a much more targeted and effective piece of content than it could have with a high-level prompt. It's like giving a human writer a complete creative brief before they start drafting.

Of course, you don't always have to wait for the AI to ask you for more details. As you get more familiar with the kinds of information different tasks require, you can start proactively including those specifics in your initial prompts. 

But collaborating with the AI to figure out what context it needs is a great way to train your own thinking and make your instructions to the AI as clear and comprehensive as possible. Over time, this back-and-forth will start to feel like second nature, and the AI's first drafts will get closer and closer to the mark.

#4: How to Connect Your AI Assistant With Other Tools

Once you've mastered prompting your AI and using its outputs in your work, you can start exploring ways to integrate it with other tools to automate even more of your workflow. This is where the real magic happens.

For example, imagine you run a YouTube channel where you regularly publish interviews with subject matter experts. Manually turning each recorded interview into a polished piece of content is a time-consuming process involving tasks like:

• Transcribing the audio into text
• Editing the transcript for clarity and flow 
• Writing a blog post version of the interview
• Pulling out key quotes or insights for social media
• Creating a video thumbnail image
• Writing titles, descriptions, and tags for the YouTube video

Traditionally, you or a human assistant would need to complete each of these steps before you could publish and promote the interview. But with a smart AI workflow, you could automate almost the entire process from start to finish.

Here's how it might work:

• You record the interview and save the audio file in your cloud storage drive to a specific folder.

• An automation tool like Zapier detects the new file and sends it to an AI speech-to-text service to generate a transcript. 
• Zapier then sends the transcript to your AI writing assistant with a prompt like “Please edit this interview transcript for brevity and clarity, organizing it with section headings and a table of contents.”
• The edited transcript gets sent to a different prompt asking the AI to “Turn this transcript into a 1,000-word blog post, including an introduction and conclusion paragraph.”
• The blog post draft gets sent to another prompt: “Please suggest 5 engaging titles for this blog post and 10 quotes that would make good social media posts or graphics.” 
• The AI-generated titles, quotes, and blog post get sent back to you for approval and any final edits.
• Once approved, Zapier automatically publishes the blog post to your website and adds the suggested social content to your scheduling tool for promotion.
• Meanwhile, the original audio file gets sent to an AI video editing tool, where a preset template adds captions, transitions, and your branded intro/outro clips to turn it into a finished YouTube video.
• The video automatically uploads to your YouTube channel with a description, tags, and other metadata pulled from your blog post and social content.

This entire process, which might have taken hours or even days of human effort before, could be completed in the background while you focus on other things—or while you sleep! And all it took was a few smart automations to connect your AI tools in a cohesive workflow.

Of course, this is just one hypothetical example—the possible combinations and use cases for connecting AI with other tools are virtually endless. You could build similar workflows for things like:

• Automatically generating reports or presentations from raw data
• Turning meeting notes into task lists and calendar events
• Drafting personalized emails or messages for each of your sales leads
• Creating product descriptions and ad copy from a spreadsheet of features

The key is to examine your existing processes for steps that could be automated or enhanced with AI. Whenever you find yourself performing repetitive, data-driven tasks that follow a predictable pattern, there's likely an opportunity to streamline things with a smart integration.

This does require some up-front investment to set up the proper workflows and train your AI with the right prompts. But that one-time effort can pay huge dividends in the long run in the form of hours saved and output multiplied. 

By connecting your AI with other tools, you're essentially creating a custom software solution tailored to your exact needs and preferences. No off-the-shelf tool can match that level of specificity and integration. 

And as AI models continue to evolve and expand their capabilities, the ceiling for what you can automate will only keep rising. The workflows you create today could be just the beginning of a fully autonomous, AI-powered work environment in the not-too-distant future.

The Future of AI Assistants

Of course, the idea of AI taking over more and more of our daily work can be both exciting and intimidating. On the one hand, it's thrilling to imagine a world where intelligent machines can handle our most tedious and time-consuming tasks, freeing us up to focus on the things that really matter. 

But it also raises concerns about job displacement, the ethics of relying on AI for important decisions, and the potential risks of building systems we don't fully understand or control.

These are all valid considerations that society will need to grapple with as AI continues to advance. But for individual knowledge workers looking to harness the power of AI in their jobs today, the key is to stay focused on the practical benefits and opportunities this technology unlocks.

In the short term, AI assistants will not replace the need for human expertise, creativity, and judgment. Rather, they will become an increasingly essential tool for amplifying and extending those human capabilities.  

The most successful professionals in the age of AI will be those who learn to work alongside these tools as collaborators and coaches. They'll be the ones who can articulate their needs and goals in ways the AI can understand and act upon. They'll be the ones who can review and refine the AI's work to ensure it meets the highest standards. And they can spot new opportunities to automate and optimize their processes as the technology progresses.

In many ways, we can think of the future of work as a partnership between humans and machines. Each party brings its own unique strengths to the table. Humans have the vision, empathy, and context—the “why” behind the work. Machines have the raw processing power, tireless consistency, and ability to instantly access and analyze vast amounts of information.

Together, they can achieve things that neither could on their own. Humans can define the destination; machines can help us get there faster. Humans can dream up the questions; machines can help us find the answers. Humans can set the standards; machines can help us meet them at scale.

This collaborative relationship will only deepen and expand as AI assistants evolve from simple tools to autonomous agents. Imagine a future where you can simply hand off a high-level goal or objective to your AI, and it can intelligently coordinate a whole team of human and machine resources to get it done.

For example, suppose your goal is to create a comprehensive report on the state of your industry for your executive team. Instead of spending weeks gathering specific data, analyzing it, and compiling your findings into a presentation, you could simply tell your AI something like:

“Please create a 20-page report on the current state and future outlook of the widgets industry. Include sections on market size and growth, key players and market share, emerging trends and technologies, regulatory issues, and strategic recommendations for our company. Support your findings with data from trusted industry sources and quotes from at least 5 expert interviews. The tone should be professional but engaging, and the visuals should align with our brand guidelines. Please have a complete draft ready for my review by next Friday.”

Your AI would then spring into action, breaking down this complex request into sub-tasks and milestones. It might start by scouring its knowledge base for existing research and data on the widgets market. Then, it could reach out to schedule interviews with industry thought leaders and generate a list of targeted questions to ask them based on the gaps it identified in its initial research.

The AI would start synthesizing its findings into key themes and insights as it gathered more information. It could create data visualizations to illustrate important trends and draft sections of the report summarizing different aspects of the industry. Along the way, it might even identify additional sources or angles to explore and adjust its plan accordingly.

Throughout this process, the AI would proactively update you on its progress and flag any decisions or issues requiring your input. But for the most part, it could autonomously manage the project from start to finish, tapping into a wide range of skills and resources to get the job done.

And at the end of the week, you'd have a polished, professional report to present to your leadership—without having to spend hours hunched over spreadsheets or slides yourself.

This is just one example of how AI assistants could revolutionize knowledge work in the coming years. By taking on more of the legwork involved in research, analysis, writing, and project management, they could free humans to focus on the higher-level thinking and decision-making that machines can't replicate—at least not yet.

Of course, this future won't arrive overnight. There are still many technical, ethical, and organizational challenges to overcome before we see AI assistants operating at this level of sophistication and autonomy. And even when they do, they will never fully replace the need for human leadership, creativity, and accountability.

However, the benefits of collaborating with increasingly intelligent machines are already clear. And the sooner we start exploring and embracing these tools, the better positioned we'll be to thrive in the AI-driven world that's quickly taking shape.

Other Notes From This Episode

• Connect with Michael Stelzner @Stelzner on Facebook and @Mike_Stelzner on X.
• Watch this interview and other exclusive content from Social Media Examiner on YouTube.

Where to subscribe: Apple Podcasts | Spotify | YouTube Music | YouTube | Amazon Music | RSS

✋🏽 If you enjoyed this episode of the AI Explored podcast, please head over to Apple Podcasts, leave a rating, write a review, and subscribe.

---

Stay Up-to-Date: Get New Marketing Articles Delivered to You!

Don't miss out on upcoming social media marketing insights and strategies! Sign up to receive notifications when we publish new articles on Social Media Examiner. Our expertly crafted content will help you stay ahead of the curve and drive results for your business. Click the link below to sign up now and receive our annual report!