What Happens to Marketing if the AI Boom Goes Bust?

There's been a lot of chatter in recent weeks around if we are in fact in an AI "bubble", with comparisons to the original dotcom bubble, the housing bubble, and their subsequent crashes. TechPolicy has an excellent podcast out that is 51 minutes of well-spent time. It's a solid exploration of the validity of those comparisons, in terms of the actual risk, the potential impact, and the important distinctions.

It's hard to argue that things aren't at least a little bit frothy at this point. AI dominates pretty much every conversation in almost every industry, whether justified or not. You pretty much can't get funded in VC circles unless you have ".ai" in your URL these days, and every marketing conference is essentially a non-stop billboard for AI adoption and promise.

But, I would argue, it's also hard to dismiss that at least in a marketing context there is real value in AI and the vast proliferation of tools flooding into our martech stacks. For all its potential downsides (AI slop, hallucination, job loss, etc.) whether it's as an brainstorming and ideation partner, a market research tool, a content production engine, or a media planning companion, AI in various flavors is proving its utility to marketers across our profession. Or at the very least, it's proving its ability to disrupt, even if many companies are still struggling to find true ROI in AI projects.

However, it's a question worth exploring. What happens to marketing if the market does in fact go "bust"? And what might that even look like? At lot depends of course on what a bust looks like in practice: an implosion in big tech and AI syock valuations is vastly different than, for example, OpenAI going bust and taking down the vast ChatGPT ecosystem with it. However, for the sake of discussion:

• Marketing budgets could get thrown into chaos: Any kind of radical shift in availability or economics of marketing AI tools and platforms could have significant impact on your MarTech budgets. What happens if to make AI economics work vendors have to dramatically spike end-user and enterprise licensing costs in order to cover the cost of compute?
• AI-dependent marketing orgs and teams could suffer paralysis: Depending on the nature of the bust, more advanced AI-centric marketing organizations could certainly be more exposed to disruption. Disruption in agentic AI solutions, in established internal workflows and employee training/expertise, and so on. If some of your major AI vendors suddenly went offline, how quickly could you pivot and how much disruption could your team weather? At what cost?
• Your marketing AI vendor mix could radically shift: To that point, the raw proliferation of marketing AI vendors is one of the biggest weaknesses in the ecosystem, and one of the most obvious echoes of the DotCom implosion. Remember when slapping ".com" at the end of anything was the path to instant riches? We're pretty much there with AI, and if the underlying economics suddenly and radically shift, we could see a vast shakeout of startups and a retreat to the major vendors. How exposed are you to AI tools and platforms from startups sitting on shakier financial foundations?

What can you do for some future-proofing?

1. Document the hell out of your marketing AI infrastructure. Your processes, tools, integrations, prompts, etc. Lay it all out to ensure vendor-portability to the maximum extent possible.
2. Develop generalized AI skillsets. Double-down on broad-based AI fluency among your marketers, vs. going all-in on the specifics of any one platform or toolset. Ensure they are ready to shift to rapidly pivot to surviving vendors.
3. Build redundancy for critical workflows and processes. Ensure you have non-AI backups, especially if you have bet heavily on smaller, early-stage startups as critical parts of your AI martech stack.

Again, give the TechPolicy podcast a listen, and take a few moments away from the never-ending sprint most CMO's are on to rapidly deploy AI, and start thinking about your contingency plans.