Read AI’s New ‘Digital Twin’ Ada Wants to Handle Your Email While You

When David Shim founded Read AI in 2021, the pitch was straightforward: use artificial intelligence to transcribe and summarize meetings so knowledge workers wouldn’t have to take notes. Four years and millions of meetings later, the company is making a much bigger bet—that AI can do more than document your work. It can actually do some of it for you.

“The way I describe our solution is that when you are bringing on a new employee, you train them. When you add Ada to your workflow and connect more services to give more context, it starts to ramp up and handle more tasks for you.” — David Shim, Read AI CEO

A Digital Twin in Your Inbox

On Thursday, Read AI launched Ada, an email-based AI assistant the company calls a “digital twin.” Unlike the chatbots that have flooded the market over the past two years, Ada doesn’t wait for you to visit an interface. It lives in your email threads, ready to step in when you CC it on a conversation.

The assistant’s capabilities fall into three categories. First, Ada handles scheduling logistics—scanning your calendar and proposing meeting times to other participants, then negotiating alternatives if the first options don’t work. Second, it answers questions by drawing on your company’s knowledge base, previous meeting transcripts, and public web searches. Third, it drafts replies to routine inquiries, waiting for your approval before sending.

Getting started requires nothing more than sending an email to ada@read.ai with the subject line “Get me started.” From there, users configure which services Ada can access and what boundaries to observe.

The Knowledge Graph Advantage

What distinguishes Ada from the growing field of AI email assistants is its foundation. While competitors rely on Model Context Protocols (MCPs)—the emerging technical standard for connecting AI tools to external services—Read AI has built a proprietary knowledge graph that maps relationships between meeting content, calendar events, and company documentation.

Justin Farris, Read AI’s VP of Product, explained the architectural choice: “Instead of just connecting to APIs and pulling raw data, we’ve built a system that understands context across your work history.” This means Ada doesn’t just know that you have a meeting at 2 PM—it knows what was discussed in your last three meetings with the same attendee, what action items remain open, and how this conversation fits into broader project timelines.

Privacy controls are built into the design. Ada won’t reveal sensitive information without explicit permission, and when handling scheduling, it exposes availability windows without disclosing the nature of competing commitments.

From Notetaker to Agent

The launch represents a significant expansion for Read AI, which built its business on meeting transcription and analysis. The company’s original product used AI to generate summaries, highlight key decisions, and track follow-up items—valuable capabilities that nonetheless kept the human firmly in the driver’s seat.

Ada crosses into agentic territory. Rather than simply providing information for you to act upon, the assistant can take action itself—replying to emails, proposing meeting times, and eventually (according to the company’s roadmap) executing follow-up tasks without prompting.

“If you mentioned a follow-up item in a meeting, Ada will ask you to set that up after the meeting with contextual data.” — Justin Farris, Read AI VP of Product

This shift from passive assistant to active agent reflects a broader trend in enterprise AI. Companies like Anthropic, OpenAI, and Google are all racing to deploy systems that can perform multi-step tasks autonomously. The bet is that busy professionals will happily delegate routine administrative work to AI, freeing time for higher-value activities.

The Competitive Landscape

Read AI enters a crowded market. Microsoft’s Copilot already integrates with Outlook to handle scheduling and email drafting. Google’s Gemini can summarize threads and suggest replies. Startups like Superhuman and Shortwave have built AI-native email experiences from the ground up.

What Read AI hopes will differentiate Ada is its deep integration with the meeting layer of enterprise work. Most email assistants treat scheduling as a calendar problem. Read AI treats it as a context problem—understanding not just when you’re free, but why certain times work better than others based on your workflow patterns.

The company also benefits from its existing customer base. Organizations already using Read AI for meeting transcription have months or years of conversation data that Ada can reference. This historical context gives the assistant a head start in understanding company-specific terminology, recurring project structures, and interpersonal dynamics.

What’s Next for AI Agents

The launch of Ada comes at a pivotal moment for AI agents. In recent months, Anthropic’s Claude Code has demonstrated that AI can handle complex software engineering tasks. OpenAI’s Operator can navigate websites and complete purchases. Google’s Project Mariner promises similar capabilities for Chrome users.

Read AI’s bet is narrower but potentially more immediately useful: focus on the communication layer where knowledge workers already spend hours each day. Rather than trying to automate complex creative or analytical tasks, Ada targets the administrative overhead that fills the gaps between substantive work.

Whether users will trust an AI to speak on their behalf remains an open question. Email carries social and professional stakes that meeting notes do not. A poorly worded reply or a scheduling mistake can damage relationships. Read AI is betting that its knowledge graph and approval workflows can provide enough safety rails to earn that trust.

For now, Ada is available to all Read AI users at no additional cost. The company plans to add premium capabilities over time, including deeper integrations with enterprise systems and more sophisticated workflow automation.

This article was reported by the ArtificialDaily editorial team. For more information, visit TechCrunch.