You know how everyone is always stressed about the holidays and all the "what do you do?" questions. I used to hate those because whether I was a talent agent, an engineer, or building brands nobody ever understood so I "put people on TV" or "I built websites" or "I made commercials." Those explanations seemed to hold everyone over which was great. But in recent years something changed... I got old.

Not actually old but you see there's a new generation of family members entering the workforce so nobody cares about me anymore—they have a second chance to get it right!

I have a cousin, he's a senior in college. Everyone is telling him what to do: become a lawyer, go to business school, someone suggested dentist (I don't even know where to begin with that one). Apparently we were taking turns because eventually everyone turned to me. My response was "do whatever you want just make sure you learn about AI." For the first time, he listened.

Thing is, whether you're in college or a seasoned marketing exec, we're all feeling a ton of pressure to build AI into our workflows. As a result of that pressure, AI is being set up as some sort of savior (which it's not). This feeling makes the whole concept of artificial intelligence seem way more intimidating than it should be. The reality is getting started is not that hard.

What is Artificial Intelligence?

Before we get started, I wanna do a real brief version of this. There are tons of great guides out there that dive into the various concepts more in-depth. We don't need those on day one so let's just keep it simple.

First off, artificial intelligence is more than ChatGPT. It's not just one app or one type of work. AI at it's core is the ability to take an input and use it to perform a task. That second part is key. Over the last decade I've built tools and solutions using machine learning, natural language processing, voice, and generative AI, and probably others that I can't remember anymore. At one point or another all of those were considered the pinnacle of modern AI, and they were all used for different things. Think of AI as a category of computer processing rather than a specific thing.

So really, it's about how you're using the output of the data you input (which could be anything). Some other key things to be aware of are:

1. Language Models (LMs) - this is the data that gets fed into the machines to teach them things. There are Large Language Models (LLMs) that are used to train more widely used consumer solutions like ChatGPT, Claude, and Perplexity. There are Small Language Models (SLMs) that are used to train proprietary solutions that maybe only your company or team uses.
2. Generative AI - this is what everyone is talking about these, essentially asking a tool to perform some kind of task for you. These days everyone is freaking out about AI generated creative work (writing, design, coding, etc.).
3. Generative Pre-Trained Transformer (GPT) - I had to look up this acronym... a GPT is a custom AI bot that you train to perform certain tasks. They require a prompt.
4. AI Agent - similar to a GPT, agents perform specific tasks but don't require a prompt. They're basically automations.
5. Prompt - this is the question you ask or task you give to an AI tool to kick off the work. When I say input above, I'm generally talking about prompts. 

There's more to it than this but nothing you need to worry about unless you start building your own apps (which I really hope you do one day).

How do I get started with AI?

Like with anything, the best way to get started is just to get started. I know it's ridiculous—you're looking for a specific process to follow—but just go to ChatGPT or Claude and start typing. It doesn't have to be about work either. Technology doesn't just have to be about efficiency or productivity, it could be fun.

So ask about dinner choices, a new workout routine, have it review your resume, or rewrite an email you're struggling with. For now, it's less about the output and more about the opportunity. Use that to ask follow-up questions and have conversations. Don't like a dinner recommendation? Let it know and tell it why. See what it comes back with. Want to use certain ingredients? Add those to your prompt. Treat it like a buddy and don't forget to say "please" and "thank you," you know... just in case of Skynet and all.

The reason I recommend just going in and winging it is because a lot of people I speak with are waiting for some perfect tool to be created for their specific need. As if some Dmitry.ai tool is magically going to appear to take on my clients and build their processes. I wish but no, it's about exploring and understanding what's possible so that when the time comes, I know where to go to get started. Then maybe one day I'll build my own Dmitry.ai that's based on these things.

That's kinda the goal, at least for me. Take the disparate pieces and figure out how they work together to make your life easier. And if they don't make your life easier, then you've lost 5-10 minutes exploring something new but now you're not wondering "what if..."

So how do you get started? Click one of the links below and start chopping it up with a robot.

What kind of tools are out there?

Taking disparate pieces is easy to do when you know what they are though and that's another blocker to using AI, there's so many tools out there now. Every piece of software on the planet is claiming they're "now with AI," so here's what I'm using these days (this list will probably change in a big way in the weeks and months ahead):

Data Analysis (ChatGPT)

For me, this is where ChatGPT shines. You can feed it a bunch of data in your prompt through links or PDFs, and ask for it to digest it all. It's nothing a few pivot tables in Excel couldn't do but who actually knows how to write pivot tables? Actually, you could ask ChatGPT to teach you in case you're curious but what's the point?

This is one of my favorite use cases for AI: taking large amounts of information and having it help distill things for you into trends, insights, and other takeaways.

Writing (Claude)

Before getting into it, I don't believe in having AI write for you. It's the Creative in me. That said, I do believe in AI giving me different ways to think about things and using that to write. In most cases, AI usually gives me just enough to make me mad and feel like I need to write it the right way myself. Whatever works, right? Ultimately in a world where you need to write the same thing 100 times to find 2-3 good options, having a bit of help is nice.

One other note here, you can easily use ChatGPT for writing too but I think Claude's writing is just higher quality.

Notetaking (Fathom)

If you've been around the newsletter, you know how much I love Fathom. It's the perfect example of what AI should do for me. It sits silently on your video calls, pays closer attention that you ever could, takes detailed notes, and then emails out the takeaways, next steps, transcript, and video. With something like Fathom, not only is notetaking automated but your next steps sync with your CRM and project management tools too so you're saving a ton of time in not having to do any of the above manually.

On top of that, it's got an "Ask Fathom" feature that will just pull up answers to questions. If you're unsure of where to start, start here.

Research (Perplexity)

Remember when you'd perform a search on Google and find yourself down some strange rabbit hole of new information? Now those rabbit holes have been replaced with sponsored results and snippets. If you're searching for directions, it's fine. If you're looking to learn, you're on your own. Tools like Perplexity have made it easier to do research. You get real results with citations and can follow any number of new rabbit holes.

It's tools like this that are changing what research will look like and that's making Google nervous.

Image Generation (DALL-E or Midjourney)

This is another one I struggle with but mostly because if you don't have a ton of time, complex image generation can be a bit of a headache. If you've never generated an image before, DALL-E (part of ChatGPT) is where it's at. Simple prompts will create images and you can edit them relatively easily. Midjourney is for more realistic compositions. They're well worth the time, in my opinion.

You can also do image generation in Canva OR you can use DALL-E to connect to Canva and then import your images directly. 

Brainstorming (Cove)

Sometimes you just need a second set of eyes on something or to bounce questions back and forth. While both ChatGPT and Claude will do that with you to a certain extent, Cove was designed specifically with brainstorming in mind (it's basically a giant whiteboard). If you're trying to figure out what the tasks you need to complete actually are, it's a great place to start ideating. 

It's also got one of the best onboarding flows I've ever seen so if you're in product, please pay attention.

Video Generation (HeyGen)

This one feels like it's out of the future but you can literally make videos without being in a video anymore. Not gonna lie, it's creepy at times but sometimes you just don't wanna be on camera but need to be. Think about video explainers or enablement materials and how quickly you can fly through them with something like this. What about personalization at scale? You ever try to record the same video 50 times? It's brutal.

This is a video I made a little while ago of robo Dmitry speaking in English, Spanish, and Russian, along with a real voiceover and robo face. At quick glance it's hard to tell that it's not me.

Video Editing (Riverside.fm)

For years video was seen as this holy grail that required specialized skills and expensive editors. Well guess what, AI has proven that that was true! You can do some amazing things with AI video editing (and I use it all the time) but it's not there just yet. That said, it's pretty cool to upload a 60-minute video and have an auto-cut feature spit out 5-6 solid edits with captions and the works.

This is the area I'm most looking forward to and companies like Capsule, and Goldcast are doing some amazing work as well.

Data Enrichment (Clay)

I can honestly say I never cared about data enrichment until I started my own business. Now I can't live without it. I can go in there and upload a list of email addresses and get names, phones, Linkedins and the like or search for specific titles within an industry and have a fully baked outreach list created in minutes.

To be honest, Clay is one of my favorite tools on this list because while many of the tools above overlap. Clay either integrates natively or through Zapier with pretty much everything (so it goes far beyond data enrichment). You could build some really powerful and fully automated "Claybooks" and "Claymations."

The list doesn't stop there. On top of these you can make slide decks, build apps and websites, create music, generate brand kits, and the list goes on. Truly, if there's a need for something there's likely an AI doing it.

Now put it all together

It wasn't that long ago that you would host a brainstorm, come up with a campaign idea, hire a copywriter, designer, and video editor to do the creative, then you'd launch and have your marketing team reach out en masse and sales team individually, all while measuring on the backend.

Now you could feed your data into ChatGPT for some insights, use Cove to brainstorm and nail down an idea, write everything from your ads to landing pages to scripts with Claude, pull your target audience list with Clay, generate personalized videos with HeyGen, and auto send everything. You can even build an agent to do it for you automatically.

The above can be way more efficient but you legitimately could set up an AI workflow that will do what took weeks and months in minutes. Am I advocating for it? Nope! Do I believe this will work? Yeah, in some cases. It goes back to understanding the opportunities and what your goals are but the point is, there's a ton of power here if you know where to look.

What about for the Brand and Creative people?

A lot of Brand and Creative folks hate AI. And I get it. They see all the generative AI and are concerned for their livelihoods. But let's be honest, have you seen most of this stuff? It's mediocre at best.

The major downside to doing what I described above is that you may lose what makes you (and your brand) unique. When you're using LLMs that are basically just giant composites of best practices everyone's work will look the same. Honestly, this is why I've never been more bullish on Creative. In this push to use AI, it's the brands that continue to differentiate that will stand out. So many marketing teams are using AI as a volume play but they're getting tuned out. Quality is more important than ever.

As Brand and Creative leaders, we should see this as a huge opportunity. So much of the work we used to get bogged down in can be automated now. All of that busy work that got in the way, can go away. 

Think about how long it used to take to create subtitles and translate them or resize images or dig through random data. It's not that these things aren't important (they're exceptionally important) but that doesn't make them a good use of your time. 

If you've ever wanted to prove that you understand how the business works, this is where you can lead the charge on self-service and automation. Find the things that your company invests time and energy in and figure out how to get them off your plate through AI (that you've trained yourself on your brand). Then use that time and energy to blow out the next big project. Don't fight this revolution, lead it.

Never outsource your thinking

Hopefully this is giving you the courage to at least give this AI thing a shot. Maybe it changes the way you work, maybe it gives you an idea of where to go next time you're unsure of what to do. In either case, my challenge to you is to make sure you keep thinking.

I've always believed that the role of AI is to do your dirty work. Let it take the menial or time consuming tasks and make them easier. Don't let it come up with ideas for you. The biggest advantage you have as a human being is your lived experience, your emotional reaction to things, and the everyday context you pick up by just looking around. That's the stuff that makes the work special. Be thoughtful and creative, then outsource the junk.

If this is the first time you're hearing about some of these tools, don't feel bad. It's never been easier to build new tools so new tools are popping up all time. The good news is that it means if you have a specific problem, there likely is a solution.

In the past few months, I've built processes, automations, and workflows using AI for enterprise B2B SaaS companies, small events companies, and even other agencies. Get in touch to get the new year started on the right foot