Adopting GenAI with Purpose — MKE DMC Launch Event Recap

Last night (March 13, 2024), we kicked off IRL meetings at MKE DMC. Below is a recap of the main discussion of the night. You’ll find a link to download the deck way at the bottom. Thanks for coming, and enjoy!

GenAI is not an easy solution for most marketing challenges. Thoughtful and strategic use requires a ton of work. So, we shared our experiences, the challenges we faced, and the methodologies we've developed to thoughtfully integrate (and test feasibility of the integration of) GenAI into our work. Here’s what we covered:

There is no one thing called “AI”

We started by talking about what GenAI is and isn't. The landscape of AI is huge and much more than the GenAI tech that’s most often covered in media and social discourse. 

The many layers include infrastructure (like data lakes and warehouses), analytics, data science platforms, and a ton of applications across industries (e.g. finance, healthcare, transportation, etc.). A tiny piece of the landscape is Machine Learning (ML) & GenAI. ML and GenAI are applied to a wide range of functions from customer experience to automation and operations optimization.

The point is, despite the significant attention GenAI receives, Large Language Models (LLMs) represent a small piece of the “AI” ecosystem. And by themselves, LLMs aren’t particularly helpful. 

The best use cases of GenAI in marketing, product, and operations includes integration with other technologies, like: data labeling, MLOps, data governance, computer vision, natural language processing (NLP), etc.

But, as we discussed, those technologies aren’t accessible to most companies. But when they are, GenAI will be the perfect user-friendly interface to the systems.

GenAI FOMO

The buzz around GenAI leads to chaotic fear of missing out (FOMO), driving reactive adoption without a strategic foundation. 

GenAI Strategic Maturity Model

We outlined a strategic maturity model for GenAI adoption based on an old SEO capabilities model that Heather Physioc introduced in 2018.

‍

Chaos

At the onset, there's 'Chaos'—when GenAI use is sporadic, without direction, and often, there's no human in the loop to guide or oversee the process. It's characterized by a lack of standards or process, where new programs and exploration happen without a solid testing framework. It's the most basic use of GenAI application, often resulting in 'one-shot prompts' (MEGA PROMPTS) that may be spammy or even harmful due to the absence of thoughtful implementation.

Absent

'Absent' is where GenAI is neither reactive nor strategic. It's pretty much not on the radar in any cohesive way. This phase is marked by a lack of direction and support for GenAI initiatives. It’s okay to be absent and you’ll go through this phase!

Tactical

‘Tactical’ is when GenAI begins to take shape as a reactive or isolated tactic. It's used, but not in a standardized or cohesive way across the organization. There's some recognition of its value, but not enough to weave it into the standard operating procedures or long-term strategy. The thing is you need to be tactical to understand the strengths and weaknesses of GenAI, and it’s part of the growing process.

Strategic

Then we arrive at 'Strategic' maturity. At this level, there's an alignment of GenAI with the organization's value system. There's transparency about how and why GenAI is used, and it's integrated well into operations with best practices in place. It’s no longer a disjointed set of tools but a cohesive part of our strategy across the organization, and in marketing.

Moving to strategic

Moving from ‘Tactical’ to ‘Strategic’ can be split into three phases: Discover, Learn, and Culture. This model is about finding practical use cases, creating an experimentation framework, and defining success metrics that align with clear business objectives. It also emphasizes cultivating a culture that values transparency and collaboration, which are foundational elements of sustainable integration of GenAI into our processes.

‍

‍

‍

‍

We noted that we have not yet seen any good examples of the next two phases.

Practice

The 'Practice' phase is where GenAI use is expected and, in many ways, mandatory for staying competitive. It involves pre-planned, advanced, and custom implementations of GenAI. There's a focus on making these tools accessible to non-technical team members, and feedback loops are critical to driving action. Here, we're not just using GenAI; we're continuously refining how we use it through best practices, training, testing, and learning.

Culture

Finally, the 'Culture' phase is where GenAI becomes ingrained in the marketing DNA of the organization. It's backed by dedicated resources, processes, and documentation. Everyone is knowledgeable and committed to continuous learning. GenAI programs are continually reviewed, optimized, and evolving. It's at this stage that we're actively seeking cutting-edge GenAI initiatives to test and implement. Our approach to GenAI becomes proactive rather than reactive. We foster a culture where every team member is encouraged to think about how GenAI can contribute to their work, ensuring that its adoption is strategic and purposeful.

structured into three phases: Discover, Learn & Iterate, and Culture. This model is about finding practical use cases, creating an experimentation framework, and defining success metrics that align with clear business objectives. It’s also about cultivating a culture that values transparency and collaboration, setting the stage for sustainable integration of GenAI into our processes.

Learning from Experience

We shared examples from our own attempts to incorporate GenAI, including Steve’s competitive analysis and Tyler’s collaborative product data optimization. The takeaway from these examples was that they only worked because we already had very specific processes mapped out. We were patient. GenAI couldn’t do everything, but it could do some of the grunt work. In both examples, we saved a little time and were able to uncover some additional insights.

The examples also supported our statement that effective GenAI use is a combination of clean data/knowledge, specific prompts, and feedback loops with GenAI.

Key Learnings

1. GenAI is really good at some things / really bad at others.
2. GenAI can leverage standardized data / knowledge repeatedly.
3. Data + Prompt + Feedback Loop (human) = Effective GenAI

Key GenAI Gaps

1. Hard to automate & repeat at scale
2. Can’t learn & adapt from feedback (outside of context window)
3. As marketers, we wish we had an accessible option for data warehousing

Looking Ahead

The future of GenAI in our industries has the potential to be promising, given there is greater accessibility and integration to other technologies like Machine Learning, Data Warehousing, etc. To prepare for this means we need to engage with the technology now, learn through doing, and develop feedback loops that help refine and improve our approaches as well as learn what GenAI is good at and what it’s no so good at.

Gratitude

Before I wrap up, I want to extend a heartfelt thank you to:

• Last night’s event sponsor, McDill
• The incredible volunteers at MKE DMC - Jaime, Emily, Grant, Casey, Lauren, Chris, Mark and Mike. We wouldn’t have been able to launch MKE DMC without them. 
• Each and every one of the 60+ people that stood with us the whole time - we WILL have more chairs next time 🙂

The energy and dedication of everyone involved are what make these conversations meaningful and impactful. Thank you for providing us with the platform to share our experiences and for helping us create a community eager to learn and grow together.

Next Steps

As we move forward, I encourage everyone to engage with the concepts we shared discussed in their own contexts. Here are a few actionable next steps:

• Reflect on your GenAI readiness
  Think about your current processes and where GenAI can add value. Are you equipped with the clean, structured data that GenAI needs? Is there a clear understanding within your team of what GenAI can and cannot do?
• Experiment thoughtfully
  Start small with controlled experiments to understand how GenAI can serve your specific needs. Use these experiments to refine your approach and build a foundation for larger scale implementations.
• Join the conversation
  Engage with the people of the MKE DMC community who you met last night to exchange insights, challenges, and successes. Collaboration and community is a powerful tool!

The next MKE DMC event on April 17, 2024, will feature discussions on topical authority with Greg Bernhardt and organic social strategies with McDill. 

Our goal is to continue to explore the intersections of technology and marketing, and we hope to see you there.

Thanks again for a perfect MKE DMC launch event!

Resources

• ‍Get access to the full deck on Canva here.
• Here's a link to the custom GPT we built for you to start your discovery.