How far are we from an AI-first tech stack?
How far are we from an AI-first tech stack?
Is it a reality yet?
The AI hype train has not only left the station, it’s gained momentum and is moving ahead at something bordering on ludicrous speed.
Let’s try and find some signal in all of the noise - what’s the reality when it comes to AI tools?
What impact could AI have on a GTM tech stack today?
In my efforts to keep pace with everything going on, I went through the exercise of trying to build an AI-first tech stack, using tools from common categories that Marketing Operations teams have tech in already today. I’ll walk you through my approach below.
But, before we get started, we need some guiding principles - we’re going to go about this strategically. So, what are the goals of an AI-first tech stack?
Don’t forget the customer (or lead)!
The first goal should always be to offer an improved experience for leads and customers. This is true regardless of the type of work you’re doing. If you can keep this as your north star, then you should always be in alignment with team goals and have a better chance to impact them positively.
As with anything new, but particularly with any new tool, it has to add significant value for the additional management overhead to be worth it.
Improvement at the speed of “prompts”
AI promises real efficiency gains in today’s “do more with less” reality. Before tools like ChatGPT were generally available, creating efficiencies required significant time and effort to be invested by the Marketing Operations team. This typically came in the form of building integrations and automation.
With AI models now able to contextualize and analyze data and take action based on it, much of the preparation needed to reach a new efficiency level no longer exists. Task completion is only a well-crafted paragraph away. Teams that can identify bottlenecks and understand how to resolve them effectively with AI should see measurable improvement in how long it takes to complete assisted tasks.
Innovation and optimization
Operations teams typically have the best view into both the tools and technology available to a team and the areas where the team has a significant need to improve.
Driving optimization through innovation helps keep the entire team looped in on new tech trends and possibilities, laying the foundation for new possibilities in audience building, targeting, and campaign execution. This effect can be magnified when focused on existing problem areas.
Finally do something impactful with all of your data
Typically, marketing teams run into three different types of data challenges:
The tools and processes aren’t set up correctly to collect usable data
The team lacks the data literacy to understand the data and what changes to make based on what they see
There is so much data that it’s impossible to know what is valuable and/or how to get value from it in any sort of scalable, repeatable approach
I believe that at this point tools that leverage AI capabilities to synthesize data and make it actionable in the tools are more useful than data-first tools that highlight insights or make it easier to comb through data.
I think we’ll see more parity here in the future, but a tool that can say “Here are the best times to email your leads based on their previous activity, click here to schedule the emails” has more utility than a tool that says “let’s help you find the best times to email your leads”.
With that, let’s get to the tools.
To be AI-first, artificial intelligence needs to account for a significant amount of a tool’s functionality. Bolt-on or piecemeal solutions that are an afterthought don’t cut it - the tool needs to demonstrate that AI is a core part of the tool’s capabilities.
I’ll summarize each category, list some potential tool candidates, and then rate the “AI Promise Realized” of the category on a scale of 1-5 robots (🤖).
For the purposes of this post, I won’t be calling out the generally available functionality in the broadly used “AIs” like ChatGPT, Google Gemini, and Bing Co-pilot. These tools are capable of a variety of functionalities, but exploring how to deploy them for a Marketing team is a subject for another post. 🧐
Content
Summary: Content creation (or “generative AI”) is the area where most Marketing teams are already actively using a tool or including some level of AI functionality in their workflows.
Tools to consider:
Written content
Images
Video
AI Promise Realized: 🤖🤖🤖🤖
Ranging from SEO tools to video creation, subject line optimization to chatbots and everything in between, generating words, images, and videos has got to be one of the most widely used and well-known use cases for Marketing teams.
You’ll likely get closer to a finished product with text-only results, image and video quality is increasing rapidly, but there are still frequent instances of the uncanny valley effect.
Email/Marketing Automation
Summary: Rather than focusing on the content creation piece (which is covered in more detail above), this section is focused on the execution of marketing automation campaigns.
Tools to consider:
AI Promise Realized: 🤖
Most of the tools/functionality here are focused on content creation and generation. There isn’t a lot of focus on audience creation, email sending, or including dynamic content from sources outside of your marketing automation or email platform.
Paid Advertising
Summary: The most effective campaign execution is nothing without the content and offers to power it, but in this section, we’re focusing exclusively on the mechanics of running an SEM or display advertising campaign. This could include things like audience creation and segmentation, category/keyword creation, ad copy/design iteration, ad groups, bidding and budget allocation, schedules/day-parting, etc.
It could also include the ability to customize the post-click experience with relevant messaging on landing pages as well.
Tools to consider:
AI Promise Realized: 🤖🤖
While there are some tools focused specifically on the campaign creation aspect of the work, it feels like there are still some big opportunities to streamline the end-to-end process, including the post-click experience, tracking, and reporting.
Social Media
Summary: The big, low-hanging fruit here is content creation. Because that is covered in the content section above, this category focuses more on the mechanics of when to post, on what networks, and controlling what gets posted.
Tools to consider:
AI Promise Realized: 🤖
Not a lot going on here, outside of content creation and repurposing existing content (i.e. pulling clips from a YouTube video for TikTok). I don’t know if there’s as big of a use case here as there are in some of the other areas.
“Real” social media is still a very human place, so the development of other tools may be slower here.
Influencer Marketing
Summary: I thought about including CRM and Marketing Automation tools in the same category, but ultimately decided to split them up because of two key points:
Most companies have two separate tools for these purposes
The main use cases for each tool are
Tools to consider:
AI Promise Realized: 🤖🤖
A couple of use cases I hadn’t anticipated include the creation of virtual influencer avatars, and the verification of influencer metrics (follower count, engagement, etc.)
The influencer marketing world is still heavily skewed for B2C, and so many of the available tools reflect that. As more B2B companies leverage influencers, I imagine we’ll end up seeing more relevant tools here.
Analytics/Reporting
Summary: This is an area that I’m very interested in watching as it evolves over the next few years. Data literacy is a struggle for many marketing teams, and without a guide to highlight what is important (and what the next best action should be), there is a significant amount of untapped value in marketing data sets.
Tools to consider:
AI Promise Realized: 🤖
Most of the functionality today centers around using natural language processing to query your data (so you can ask things like “how many visitors requested a demo on our website in the last 30 days?”).
This is a helpful feature, but I’m hopeful that we see a lot more related to the “why” behind the story the data is telling us. That seems like a “true” AI use case - look at this mountain of data and tell me what I should care about and why.
SDR/Inbound
Summary: Next to content creation, this use case has the most momentum. There. area variety of big players in this space, a lot of people paying attention to what’s going on, and lots of overlapping functionality.
Tools to consider:
AI Promise Realized: 🤖🤖🤖🤖
Even if you set aside the email personalization functionality, this is probably one of the more well-supported Marketing use cases. Functionality exists to identify potential prospects, customize chat and email messaging aligned with their specific details, execute outbound outreach, and even analyze and surface insights from phone and email conversations.
Wrap Up
In reality, we’re definitely not to the point where an AI-first tech stack is a practical reality, but with the speed at which AI innovation is accelerating, the gap between reality and possibility is getting smaller daily.
The goal of this post was to explore the possibility of building an AI-first tech stack. This will be interesting to revisit in a year (or even just six months) from now, to see how the feasibility has changed. This post could very well be out of date in a month, as meaningfully different and capable tools are being announced weekly. Perhaps this post will be revisited in the future.
If you’re interested in things that are more realistic to try today,
explored some of the current top AI use cases for Marketing in a recent post on his RevOps FM newsletter, which I’ve linked here:
His recommendation of taking a use-case-first approach is something that is much more achievable now and is a best practice when it comes to evaluating any new tool, let alone AI-specific ones. I’d highly recommend taking a look. He has compiled an Airtable with 35+ use cases to review as well.
Like it or not, AI is here in all of its buzzwordy gloriousness. Whether you’re just keeping tabs on how it can be used, or you’re actively trying to incorporate it into your workflow, it is definitely past the “pay no mind” point. We could be on the cusp of a pseudo-AI revolution, or this could just be the next noticeable wave in the tech space.
Either way, taking a proactive approach will certainly be beneficial to your career and better position you to take advantage of the benefits AI can (and will) bring.