Deep: AI Agents in Practice
How companies are using AI Agents in the real world. Plus new features, SDKs and applications of AI agents for inspiration.
🔒DoP Deep goes deeper into the concepts and ideas that are covered in the Weekly Briefing to help you learn lessons from top tech companies. If you’d like to upgrade to receive these in-depth pieces of analysis you can upgrade below. New reports added every month.
Despite Google’s VP of product Ryan Salva declaring that he “hates” the term, the hype train for AI agents is still very much in motion.
To be fair to Mr Salva, he's not alone in his criticism; Andrew Ng also recently shared his dismay at the term, and part of the reason it's ruffled a few feathers is that technical people love strict technical definitions. This works for things like APIs and MCPs, but when it comes to "agents," the waters get a little murkier. Instead of a shared technical definition, companies have stretched the label "AI agents" to fit a wide range of capabilities, often with little consistency.
In this Deep Dive, we’ll attempt to figure out what the term AI agents actually means by looking at how exactly companies are using AI agents in the real world along with the latest new AI agent releases that can be used by product teams and an exploration of perspectives from real world tech leaders.
These real world examples include 20+ new AI agent features from leading companies like Adobe, Hubspot, Visa, Replit, Wix and others as well as new AI Agent startups and SDKs that have recently been released for product teams to build their own AI agent features.
If you’re feeling a bit jaded by the term AI agents but you’re still keen to stay in the loop with how they’re being used, this deep dive should hopefully help.
Coming up:
How companies are using AI agents internally for things like release notes, financial disputes, product development and more.
The big bets being made by startups like Manus and Genspark, and why they think agents can replace entire workflows
How one company managed to redeploy 2,000 staff members elsewhere after AI agents were used to do their jobs for them
Insights from Grammarly’s CEO, leadership at Salesforce and Databricks on where AI agents are headed in the future
New AI agent SDKs: what they are and what they do
The full list of 22+ AI agent examples and releases
How this Deep dive report is structured
The deep dive includes 20+ different examples of AI agents in action. This includes examples of how companies are using agents internally, new AI Agent features recently released plus new agent development kits and SDKs.
Company - the name of the company featured. This includes a mix of larger companies (OpenAI, Google, Microsoft) and new startups like Manus, Cognition and Genspark.
AI Agent example - the name of the AI agent or the product e.g. Amazon’s “Buy for Me” agent, Cognition’s “Devin”, Wix’s “Astro”.
How it works - a summary of how this agent works and the value it offers to users. The report includes examples of how companies are using AI agents in their own companies, new AI agent features that have been recently released and new frameworks and SDKs that product teams can use.
Category - this Deep dive covers a wide variety of different categories - more details on that below.
Model - where possible, the AI models that are used in development of the Agent. This includes GPT-4o, Anthropic’s Claude, Gemini, Llama and proprietary models.
The AI Agent categories explained
The report includes a mix of different product categories including:
Banking and finance companies like Visa, Capital One and others who use AI agents across various different areas of their business.
AI products - the product itself is an AI product and the agentic capabilities are part of that product e.g. Manus.
SaaS companies like Atlasian, Hubspot, Adobe, Wix and others.
SDKs and frameworks - as well as new features which show AI agents in practice, new Agent SDKs and frameworks are included to bring
A closer look at the AI agents and how companies are using them with practical examples
Now let’s take a closer look at the examples included in the report, starting with some real world examples of how companies are using AI Agents internally.