Why GenAI Needs a “Brain”

It’s been more than a year and a half since ChatGPT launched. In that time, we’ve seen an unprecedented rapid development and adoption of generative AI (GenAI). One-third of respondents to a recent McKinsey Global Survey of C-suite executives are personally using GenAI regularly in at least one business function; and 40 percent of them expect to invest more in AI overall.?

This increased investment in GenAI has surely driven gains in productivity and business opportunities in recent months, but it’s also creating a bit of a mess for enterprises. The patchwork of GenAI tools that many companies have adopted across business functions can lead to internal miscommunication, lack of information sharing and even outputs that confuse customers. In this, GenAI can negatively impact efficiency and other critical business outcomes –– the very areas most businesses are aiming to improve with these solutions.

This is a tricky tradeoff.

The surge in GenAI adoption means that, across industries, organizations are now running multiple, disparate GenAI systems — and this fragmented mess of GenAI is causing problems.?

The missing link

Having to integrate myriad tools is not just an IT problem. The assortment of different tools (by channel, by task, and so on) is creating a ripple effect of inconsistent information and cross-functional miscommunications. The lack of a feedback loop for shared insights can handicap an entire enterprise, from digital teams to sales to industry partners — potentially jeopardizing growth opportunities.?

GenAI fragmentation impacts every function.

Consider this example within the marketing function. The beauty of GenAI stems from LLMs that are continuously learning, right? As businesses adopt several GenAI tools across the martech stack (54% of marketers say they’re using 50 or more martech platforms), for example, they run the risk of developing multiple brand personalities. Each tool could generate content in a different voice, such as humorous messages for email, but more conservative messages for web copy –– leading marketers to risk sending mixed messages across channels and audiences. The resulting disjointed brand voices make for a confusing customer experience.

When different teams use different systems for GenAI, there’s no baseline or consistency for cross-team knowledge sharing — no automated feedback loop to gain insights about challenges, wins, and lessons learned, and thus apply best practices and outcomes. For instance, the social team would benefit from knowing what content performed well for email campaigns, and the ecommerce team could learn from the SMS team’s successes.?

When AI capabilities are siloed or fragmented, you end up sending mixed messages to your customers across channels which can jeopardize retention, loyalty, and opportunities for upsells. For marketers, this can make or break a campaign strategy and threaten the brand voice they’ve worked so hard to build. For highly regulated industries, such as banking or pharma, the stakes are even higher.

Enter the GenAI brain?

GenAI in the enterprise needs to shift from being an output as a service tool to an intelligence as a service layer: a “brain.” The role of this brain in marketing, for example, would be to generate, evaluate, analyze all communications to enable productivity, compliance and performance across channels. It must also be connected to consumer responses to learn and evolve how a brand communicates with its customers. That feedback loop is the sensing ability of the brain; crucial for its evolution.?

Apoorv Durga, Vice President of Research & Advisory at Real Story Group explains it well in a recent post. He recommends considering AI as a foundational component in your stack. “In other words, abstract out as much of the AI functionality as possible and move it into a layer underneath your engagement platforms, even as it gets invoked and runs within individual platforms.”

?

In practice, this means teams will more easily learn from each other's wins and challenges, mitigating some of the disjointedness resulting from current fragmented approaches. With improved and consistent data and insights flowing from the GenAI brain, businesses can make more confident decisions. They will also be able to create more meaningful customer experiences across channels and drive greater revenue.?

What’s more, with increasing pressure for regulation to mitigate hallucinations, risks, and misused data, the governance framework of a GenAI “brain” provides enterprises with certainty, compliance, and customer trust.?

Fitting the pieces together

You’re probably wondering which tech vendor(s) have what it takes to be the GenAI brain. I’m wondering too. I do know the answer when it pertains to a brain for the Marketing function…

What’s clear is that it cannot be an out-of-the-box implementation of a genetic foundation model that can return non-compliant content, or irrelevant or wrong results, or non-compliant content, while also lacking the ability to understand consumer behavior. To truly control the system such that output and outcomes are aligned with strategy and regulation, the more likely solution would be a proprietary/industry specific LLM stack that can be tailored to meet an enterprise’s specific needs and streamline performance, brand voice and legal compliance across channels.

As you think about your next GenAI investment, or whether you are building your own models, it’s essential to plan for a GenAI as an “intelligence as a service” layer. I’m confident that enterprises will see their teams and outcomes become more optimized, efficient and trustworthy as a result.?