Dust Off the SWOT: How to Choose the Right AI Marketing Use Cases.

Summary

AI integration is hard not because of what AI can do, but because of choosing where to start. The fix is boundaries, and the tool is an AI-adapted SWOT: Strengths and Weaknesses surface internal use cases (enhancing a strength is usually the faster win), while Opportunities and Threats handle timing and the disruptions that derail adoption. Two client examples ground it. The takeaway: set your own boundaries, involve leadership, and plan the changes.

When I did education research, we used to examine how boundaries and structure can enhance creativity.  I am sure you have felt this: it is often harder to be creative staring at a blank slate than at a system with clear functions and boundaries.

I find AI in marketing to be dizzying in possibility, and when considering B2B use-cases I feel more like I am staring at that blank slate.  If you can think of a thing, there is someone out there saying AI can do that (probably selling a course).  Choosing a use case, let alone prioritizing them, feels futile given the speed.

And yet you must integrate AI into your org structure.  In times such as this, it is important to take a second to set the boundaries.  This will enable all sorts of creativity from your team, and ultimately yield better results.  And how better to set boundaries than to take a look inward at the skills and gaps of your team?

That’s right friends, it is time to dust off the ol’ SWOT.

A quick refresher for those who have left this foundational relic of strategic thinking in the past: Strengths, Weaknesses, Opportunities, Threats.

This is often where I start when considering marketing orchestration workflows or any kind of structural change.  SWOT is a framework, and needs to be adapted to the situation at hand, so here is how I define these with clients and the types of things I look for.

Strengths and Weaknesses

When looking at these two, I suggest a simple: what is my team good/not good at?  Keep it simple here.

These internal indicators are great for defining use cases.  I like to break things down into 2 categories:

1. Things that enhance what we are good at
2. Things that cover for where we are weak

As a team, you will need to prioritize one of those.  You can likely do multiple use-cases at once (depending on the situation), and I recommend you pursue both.  Eventually though, team demands will shift, so prioritizing one will help with decisions and clarity at those points.

As a general rule, I have found it faster and easier to go with option 1.  Weaknesses and gaps take time to cover properly, and it is likely that your foundations aren’t fully developed yet.  Those foundations are crucial to AI being able to operate, and poor fundamentals tend to congregate around team weaknesses.

In the long run, filling in important gaps will likely be more beneficial to a team with broad categories of work like Marketing.  So again, I recommend having both of these sorts of projects going, but quicker gains can be made through enhancing strengths.

Opportunities and Threats

This one takes a bit of a different approach than normal.

Opportunities = upcoming things where AI can make a visible, valuable impact

Let’s face it: you need to prove value to prove budget, so don’t just get outcomes, make sure they are visible.

Here are some examples of things I look for:

• Corporate goals/objectives
• Roadmap events (like launches and/or conferences)
• Tech stack AI updates

Opportunities define when/where you want to drive AI adoption.  It also helps you to prioritize your strength/weakness use cases.  Good targets tend to be a few months out and without grueling lead-ups.  Things can be done fast with AI, but adoption and output training still take time.

Threats = upcoming things that may threaten your ability to adopt

You still have to get work out the door, even while implementing these tools.  Threats are what get in the way of those two things happening in harmony.

Here are some examples:

• Ad-hoc requests
• Opportunistic plays
• Late-term feedback
• Corporate shakeup

These are for what “it is what it is” was first uttered.  The four horsemen of the project plan.

While you probably can’t see these coming, it is important to plan for them.  If one (or more, barf) arises then you are going to have to prioritize.

Some examples

Intake

One of my recent clients was looking to rework their marketing intake flow.

The gap:

this team had an internal agency structure with a central marketing team serving the needs of multiple business units.  As anyone who has worked in that knows: intake is crucial.  Their intake process was open-ended.

The need:

Garbage in, garbage out.  And what do you get from 5+ teams all giving input in different ways?  Garbage.

The solution:

So they embarked upon the long play: using AI tooling to standardize intakes.  The solution looked something like this:

1. Multi-format inputs into a central agent
2. Central agent uses natural language processing to format
3. Scheduling agent finds alignment time for requestor and recipient
4. Central agent passes output to both requestor and recipient
5. Alignment occurs

A simple idea, but with multiple agent teams and workflows.  Then, you had the biggest task of all: driving adoption among the disparate siloed business units, and receiving teams.

Change management is a must here.  I would love to chat about that sometime, if you want to know more about how to make this work.

Research

This one I was not involved with building, but it was an extremely interesting case.

The strength:

This team was strong in sourcing new target accounts.

The need:

They wanted to expedite their ability to enrich the account data, given that popular data providers couldn’t given that the accounts were small businesses, largely brick and mortar.

The solution:

A coded script and team of agents.  The solution looked something like this:

1. Pull public record account data and save to a folder
2. Take in accounts and what information is available
3. Chunk accounts and initiate sub-agents
4. Run sub-agents in parallel to enrich accounts
5. For all identified contacts, create hubspot objects and assign tasks to reps

Again, a simple workflow but with numerous agents and integrations behind it.  However, once built, this produces a big enhancement to a well-understood and utilized process.  It just saved someone a bunch of time.

Conclusion:

There isn’t a right way to do this.  The boundaries are pretty open, so you are going to have to make your own.  It takes some time, I would recommend getting your leadership team together and working through this.  Then source more specific use-cases to target those things from your front-liners.

Project and change management disciplines are important here too, so don’t skip those.

Anyway, if you want to talk more about this, find me at accelerate@heinzmarketing.com.