January 29, 2025

DeepSeek and the importance of a multi-modal approach

Why Stagwell Marketing Cloud stays nimble.
MarTech
TABLE OF CONTENTS

DeepSeek, a Chinese-developed AI platform that seemingly overnight became a dominant force in generative AI, has reignited fierce debates in our industry this past week.

From the perspective of Stagwell Marketing Cloud, it’s not very disruptive. DeepSeek is just another LLM (albeit one that claims to have been trained for a fraction of what larger American AI firms are spending). 

It might spell bad news for larger companies that are currently building their own LLMs and investing billions in resources. But for those creating purpose-built AI solutions, the name of the game has always been adaptability. The ability to rapidly integrate emerging AI technologies, rather than being locked into a single model or provider, is becoming a competitive advantage. 

Companies like Stagwell Marketing Cloud that take a purpose-built approach—designing solutions that remain nimble and can leverage multiple AI models—will be the ones that endure.

An “agnostic” strategy

As Lindsay Hong, CEO of SmartAssets, has noted, not all large language models (LLMs) are created equal. In a discussion about SmartAssets’ approach, she emphasized the importance of model selection, stressing her “agnostic” approach to LLMs. (Translation: Whichever tool is best for the job at that moment, gets the job.)

The reality is that different LLMs excel at different tasks. Some models may be better at creative writing, while others are stronger in reasoning, summarization, code generation, or visuals. A well-structured AI solution doesn’t rely on a single model but rather adapts dynamically for the best fit.

As Hong further said: "Eventually, everyone will have access to the same tools at the same price. The nuance is going to be how you prompt them, and which model you choose based on knowing what it’s specifically good for."

Companies that fail to recognize the strengths and weaknesses of various models risk producing shoddy results, while those that leverage a multi-modal approach (and add purpose-built functionality that eliminates the need for a trained ‘prompt engineer’) can create truly differentiated AI experiences.

Stagwell Marketing Cloud’s adaptive AI vision

At Stagwell Marketing Cloud (SMC), we have always taken a multi-modal approach to AI, recognizing that the key to long-term success isn’t allegiance to a single LLM but rather the ability to integrate and adapt to the best available models. 

Our suite of products is designed with this flexibility in mind, ensuring that as new advancements emerge—whether from DeepSeek, OpenAI, Anthropic, or other players—our AI-powered solutions can absorb and incorporate these improvements.

The DeepSeek phenomenon is a reminder that AI isn’t slowing down—it’s evolving. The companies that succeed in this space will be the ones that remain agile, embrace a multi-model strategy, and leverage the best of what AI has to offer at any given moment.

Louis Criso

Louis Criso is the Head of AI Solution Development at Stagwell Marketing Cloud.

Take five minutes to elevate your marketing POV
Twice monthly, get the latest from Into the Cloud in your inbox.
Related articles