Because LLMs are non-deterministic like humans, it's more effective to integrate them using existing human-centric processes. Give an agent an email, permissions, and "onboarding" so it can navigate the organization like an employee, rather than building complex new software interfaces.

AI won't magically fix a broken strategy. The key is to identify what already works—your best emails, responses, and processes—and use that proven data to train the agent. This approach scales your known successes rather than hoping AI will invent a winning formula from scratch.

To overcome employee fear, don't deploy a fully autonomous AI agent on day one. Instead, introduce it as a hybrid assistant within existing tools like Slack. Start with it asking questions, then suggesting actions, and only transition to full automation after the team trusts it and sees its value.

Vercel's CTO Malte Ubl suggests a simple method for finding valuable internal automation tasks: ask people, "What do you hate most about your job?" This uncovers tedious work that requires some human judgment, making it a perfect sweet spot for the capabilities of current-generation AI agents.

Don't underestimate very junior talent who are native to new AI tools. A recent Stanford grad at Laurel built a 'chief of staff' agent for the sales team, automating call prep by scraping internal and external data. This highlights a new source of high-leverage innovation.

Vercel builds internal AI agents and tools, like an Open Graph image generator, to automate tasks that were previously bottlenecks. This not only increases efficiency but also serves as a critical dogfooding process, allowing them to innovate on their core platform by building the tools their own teams need.

Sendbird created an internal platform where employees post 'quests' for AI tools. This marketplace connects needs with builders (engineers or AI-enabled staff) and even AI agents, bypassing slow prioritization processes and fostering a building culture.

For companies given a broad "AI mandate," the most tactical and immediate starting point is to create a private, internalized version of a large language model like ChatGPT. This provides a quick win by enabling employees to leverage generative AI for productivity without exposing sensitive intellectual property or code to public models.

The barrier to creating AI-powered solutions has dropped dramatically. An HR team member with no AI expertise built a Slack bot trained on the employee handbook to answer common questions, saving hours of repetitive work. Every department should be empowered to identify and automate its own low-value, repetitive tasks using accessible AI tools.