Enterprises are rapidly shifting away from generic artificial intelligence solutions toward tools deeply personalized to their users, a move driven by the demand for more relevant and efficient AI experiences. This isn’t simply about better recommendations; it’s about AI systems that understand individual workflows, terminology and priorities, ultimately boosting productivity and unlocking new capabilities. The focus is shifting to AI that learns *from* users, rather than attempting to predict their needs, a trend experts say will define the next generation of enterprise AI.
The core principle behind this shift, as articulated by Lijuan Qin, head of product at Zoom AI, is a move away from guesswork. “Don’t try to randomize, or guess who I am. I advise you, Here’s what I care about,” Qin explained in a recent Beyond the Pilot podcast. This emphasis on user-defined preferences is fueling the development of AI agents capable of analyzing individual user data to create highly customized experiences.
Zoom is at the forefront of this trend with its AI Companion, a generative assistant that goes beyond standard features like meeting summarization and action item creation. The AI Companion now tracks opinion divergence and user alignment during meetings, offering a more nuanced understanding of discussions. Users can tailor meeting summaries to their specific interests and generate targeted follow-up emails for different contacts – whether they’re salespeople or account executives – with the AI automatically populating the documents post-call. Zoom AI Studio allows for the creation of custom dictionaries to process unique enterprise terminology, ensuring more relevant AI outputs, and a deep research mode delivers analyses based on both internal expertise and external insights.
Crucially, Zoom is prioritizing user control. Qin emphasized the importance of “extremely clear controls” on agent permissions, allowing users to dictate whether the AI can automatically send emails or trigger verification steps when sensitive information is detected in transcripts. This level of control is designed to mitigate the risks associated with AI inaccuracies and ensure data security. “The most important thing is we do not assume AI is smart enough to gain everything right,” Qin stated.
This demand for personalization is driving what Sam Witteveen, co-founder of Red Dragon AI and host of the Beyond the Pilot podcast, describes as a “land grab for context.” “Definitely knowing your users is the big thing, right? Knowing what apps they are living in, what day-to-day tasks are they constantly doing?” Witteveen said. “Companies realize the more they have about you, the better the [AI] memory can get, the better they can customize.”
Several applications are emerging as leaders in this space. Witteveen highlighted Claude Cowork and OpenClaw as examples of models capable of making decisions for users based on their accumulated context. With OpenClaw, users can customize the AI to perform tasks at specific times, such as “Hey, at 4 o’clock I want you to do this.”
However, this increased personalization comes with challenges. Witteveen cautioned that token usage and security are paramount concerns. OpenClaw, in particular, has faced security issues since its launch, prompting some enterprises to uninstall the autonomous agent or ban its use entirely. Proper uninstallation procedures are critical to avoid inadvertently deleting entire enterprise stacks. The cost of personalization, driven by increased token consumption, must be carefully considered, requiring businesses to track relevant metrics.
The shift towards personalized AI also raises questions about the “build vs. Buy” dilemma for enterprise software. As AI capabilities become increasingly central to business operations, companies must decide whether to develop their own AI solutions or integrate third-party tools. The urgency of this decision is growing, with some experts suggesting that companies that fail to experiment with AI skills risk falling behind.
The development of large language models (LLMs) and AI agents is rooted in decades of progress in natural language processing (NLP) and machine learning, according to IBM. LLMs are trained on vast amounts of data, enabling them to understand and generate human language, and perform a wide range of tasks. These models utilize a transformer architecture, excelling at handling sequences of words and capturing patterns in text.
As enterprises continue to embrace personalized AI, the ability to effectively manage data privacy, security, and costs will be crucial. The next key development will likely center on refining the balance between AI autonomy and human oversight, ensuring that these powerful tools are used responsibly and ethically. The conversation around responsible AI implementation, including data governance and algorithmic transparency, is expected to intensify in the coming months.
What are your thoughts on the rise of personalized AI in the workplace? Share your experiences and insights in the comments below.
