Circleback CEO Ali Haghani: Recording Company Meetings Will Become The Norm

Circleback's Functionality and Setup [0:00]

• Circleback is an AI note-taker that records, transcribes, and organizes conversations.
• It assigns action items and allows for custom automations to update other applications.
• The tool can be queried across all team conversations, acting as a "company brain."

  "Circleback is an AI note taker that captures and organizes everything from your conversations."

Minimalist Hardware and Keyboard Choice [0:37]

• The co-founder prefers a minimal setup with few tools and desk items.
• A Topre keyboard, originally with Japanese manuals, is highlighted for its quiet, responsive typing.
• This keyboard's "early registration of each key press" allows for less forceful typing.

  "But it's quite nice to type on because it's got a very nice typing feel, but it's also not super loud and it has a very like early registration of each key press."

The Fidget Toy and AirPods Max [1:29]

• A small vehicle toy, acquired from a Temu ad, serves as a fidget item during discussions or dictation.
• AirPods Max are described as a "status" item, signaling focus similar to old instant messenger statuses.

  "It's honestly more of like a status, not in terms of like signaling to the rest of the team that like you're locked in."

Unexpected Uses of Circleback [2:10]

• Circleback is used as an Applicant Tracking System (ATS) to track outstanding action items from interviews.
• It helps manage candidate interactions, interviewers, and timelines.
• The company page for Y Combinator shows all contacts, outstanding action items, and conversation history.

  "So this view is basically any action item that belongs to someone at Circleback that is not done and it's from an interview."

Interview Preparation and Recording [4:19]

• Circleback was used to prepare for this interview by reviewing past emails and conversations.
• A press-and-hold action button is set to start Circleback recordings quickly.
• An upcoming watch app will allow starting recordings via a watch complication.

  "I use multiple times a day is just pressing and holding the action button."

AI Agents and Workflow [6:05]

• The co-founder spends 10-20% of their day in OpenClaw, with most time in coding tools.
• Agents are created for finance, customers, and people ops, with processes moving to Slack channels.
• Circleback captures customer feedback from meetings, allowing questions to be asked about specific feedback points.

  "We have my like main Claw agent and then I have one for finance, one for customers, one for people ops."

Coding Tools and AI Integration [8:06]

• The co-founder consciously reduces direct coding to avoid bottlenecks, focusing on bringing on more engineers.
• Thursdays are designated as "no meeting" days for focused work.
• AI agents are heavily used for tasks, with the bottleneck shifting to code review.

  "And then the bottleneck then becomes code review, which we've also done a lot to kind of spawn different types of agents depending on what the PR is and making sure that everything, not just from, like, a code quality standpoint, but also architecture and, like, product."

Testing and Writing Style [10:50]

• Evals are in place for quantifiable tasks, such as identifying action items.
• The goal is to make Circleback notes sound human-written and avoid redundant phrases like "discussed."
• Continuous improvement of models and prompts is crucial for maintaining quality.

  "It's one of the primary surface areas of the product."

Token Maximizing and AI Budget [11:55]

• The company culture emphasizes not just completing tasks but making them easier for future efforts.
• AI is used extensively, with an unlimited budget for AI tools to ensure high-quality output.
• Team members share new tools and usage methods during show-and-tells.

  "Right now, we have an unlimited budget for, like, using AI, and we trust the judgment of everyone on the team."

Workflow Evolution [12:45]

• The co-founder spends less time in the editor, using Zed for quick diff reviews and small changes.
• Most work is done through agents, automating operational tasks previously handled manually.
• This shift allows more focus on high-leverage activities.

  "But most of my work is, like, directly through agents when writing code, I think I'm doing a lot less of the things that I didn't like as part of being a founder, which is the ops related work where I was like, I don't need to be logging into our payroll software to be doing this."

Agent Limitations and AI Thesis [13:45]

• Agents are not trusted to send emails independently or copy product without review.
• Data access and security tasks are reviewed by agents but architected by humans.
• The thesis is that as LLMs improve, recording all conversations becomes essential for AI context.

  "The context of everything that's going on becomes super important because otherwise your AI agent is working in isolation."

Future of Software Engineering [15:23]

• The bounds of software engineering are increasingly defined by imagination rather than technical skill or bandwidth.
• As software writing costs decrease, the focus shifts to what and how to build.
• The field is rapidly evolving, making it an exciting time for builders, especially in startups.

  "And I think as the cost of writing software becomes lower and lower, what matters more is not how fast you ship, but what you decide to build and how you decide to build it in order to be able to continue composing the systems that you build together and having them interoperate and work together."