Christian Monberg, Chief Technical Officer and Head of Product at Zeta Global.
For marketers, the goal is always to find ways to make sophisticated marketing simple. Even today, in its nascent stage, GenAI is delivering exponential increases in efficiency in a game where speed is paramount. In less than 30 years, we’ve gone from basic autocorrect to ChatGPT writing novels. Bard knows more than 20 programming languages and can freely converse in English, Japanese and Korean.
While the resultant value of these tools has increased exponentially from the early days, these are still the early days. Today, we’re seeing AI being used to augment or evolve existing tools and processes, but what new tools will surface that are centered around GenAI rather than just taking it on a blind date? This is the time to take stock of where we’re at and plan for what comes next.
Current State: Feeling Out The (Temporary) Limitations
In terms of the technology landscape, we can expect to see the typical cycles of proliferation and consolidation. Microsoft is out early, but we’ve only seen the tip of the iceberg on Google’s Bard. And don’t count out Apple or Amazon, especially with proprietary models running on proprietary hardware and chipsets.
As a result of this figurative arms race, the large language models (LLMs) that make this possible are becoming a commodity, and competition will continue to drive innovation both in the core technology and in its application.
We’re already beginning to see practical applications getting integrated into marketing platforms. You can use conversational interfaces to gather consumer insights, write copy, code or generate creative. It’s easy to point out flaws: Tools are still buggy and slow, and creativity is still in the uncanny valley, but how far they’ve come is mindblowing. That they need human supervision to be effective is a feature, not a bug, and like any muscle, they will only get stronger with use.
Today, we have lots of use cases but fewer workflows. The industry hasn’t meaningfully integrated it into our platforms. You still need to know how to use the tools and layer them into existing processes. Expect this skills gap to be short-lived: As AIs get access to better data, they will produce better results, and we will shift from automating tasks to accelerating entire workstreams.
Near Future: The Institutionalization Of AI
AI That Is Integrated Into Marketing Platforms
The very next thing we’ll see are bolt-on solutions on existing platforms: AI virtual assistants and plug-in type apps for generating copy, code or creative—but they are just the beginning. We need to uplevel the thinking, from content to campaigns to experiences.
Right now, we’re seeing how GenAI tools can blast through the two biggest bottlenecks in any marketing process: analysis and content. The next step is to connect the dots between systems of insights, content creation and systems of action.
Within the next three to five years, I believe we’ll see AI integrations that are able to create and deploy whole campaigns with minimal intervention. If they’re making us 10% more efficient now, imagine what 90% looks like. This will involve a massive realignment of workflows and roles, as writers, designers, analysts and coders shift from being practitioners to being editors, ideators and co-creators.
AI That Is “Data-Aware”
Trust is one of the biggest hurdles—and rightly so. AI systems are prone to bias, inaccuracy and even hallucinations. We need to ensure the right governance is in place.
One way to attack the trust issue is through domain-based intelligence. Training LLMs in specific areas of expertise can dramatically increase the quality and relevance of their output. We’re already seeing the emergence of LLMs that are specific to industries (think finance, healthcare).
The next step will involve LLMs that can be rapidly trained on more niche areas (think “financial management software for influencers” or “your brand”). This is how GenAI content could move from passable to thought leadership.
Another way developers will address the trust issue is by improving the provenance of AI-generated output. In simple terms, this means having the AI cite its sources. If we have a better understanding of where data comes from and how it can best be used, it will make oversight more efficient and enable a tighter coupling between creating content and acting on it.
AI That Can Interact With Its Environment
You can ask ChatGPT for a casserole recipe today. What if it could also automatically inventory your fridge, search out the best local deals on missing ingredients, order the groceries and preheat your oven? Developers are starting to use the term “agentic” artificial intelligence, which simply means AI that has agency within your system. This is how a marketer could essentially input a brief and get a ready-to-deploy campaign.
As the ecosystem evolves, more solutions are hitting the market. Adept.ai is building an LLM that can “interact with everything on your computer.” Langchain is building a framework for developing applications on top of LLMs. AWS is launching Bedrock to make it easier to access these proliferating intelligences. The takeaway is that, soon, AI won’t be bolted onto some of your applications; it will be foundational to them—and to our lives.
Cultural Implications And Unknowns
This technology is incredibly powerful. The possibilities are limitless, and so are the pitfalls. Will we see it used for irresponsible marketing, malicious deepfakes and election interference? Yes, I believe so. Will we also see new protein folding techniques that unlock treatments for rare diseases? Again, yes, I believe so.
A healthy combination of optimism and skepticism will be required to navigate this new era, but AI isn’t something that’s being inflicted upon us. We have agency in shaping how it’s used, which is why we have to keep innovating—both in what we do with it and how we govern it.
Early adopters will always have an advantage, but that window is closing. Now is the time to think seriously, and creatively, about what the future of your work looks like.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?
Read the full article here
