CHAPTERS
- 0:00 – 1:00
Why Claire switched from OpenClaw to Grok Bot (and what you'll learn)
Claire sets the premise: after weeks of hands-on use, she’s shut down her OpenClaw agents and moved most workflows to Grok Bot. She previews a practical, template-driven tour of her daily bots—what’s better than before, what’s still missing, and which ideas viewers can copy.
- •Moved “almost entirely” from OpenClaw to Grok Bot after real usage
- •Episode focus: practical use cases, not theory
- •Promises bot links/templates in show notes
- •Will cover business bots first, then fun personal bots
- •Will also cover what Grok Bot still lacks
- 1:00 – 2:01
Sponsor break: WorkOS and enterprise-ready authentication/controls
A sponsor segment explains why AI tools need deep system access and why that creates security/compliance pressure in enterprise sales. WorkOS is positioned as “Stripe for enterprise features,” providing authentication, access controls, and audit logs via APIs.
- •AI apps need deep access to codebases/docs/support systems
- •Enterprise buyers demand strong security controls and auditability
- •Building auth/access/audit from scratch is costly
- •WorkOS offers drop-in enterprise APIs
- •Examples: OpenAI, Perplexity, Cursor use WorkOS
- 2:01 – 4:01
Grok Bot fundamentals: bots, plugins/connectors, and the built-in computer
Claire gives a quick product tour: Grok Bot is a multi-agent platform where you define bots per job, connect them to tools via plugins, and let them operate inside a cloud VM. She emphasizes how these primitives map well to a “bot as coworker” mindset.
- •Bots are discrete agent threads with scoped roles/context
- •Naming a bot can help Grok Bot infer its job
- •Plugins combine skills + connectors across your stack
- •Each bot has a cloud computer/VM for browsing, installs, running code
- •Routines enable scheduled, recurring work
- 4:01 – 7:03
Why Grok Bot connectors feel better (multi-account + easy OAuth)
Claire explains the moment she “fell in love” with Grok Bot: the connector experience. Bots often prompt for the right plugin at the right time, and Grok Bot supports multiple accounts per plugin—critical for founders juggling many inboxes and workspaces.
- •Bots can prompt you to connect needed plugins on demand
- •Simple OAuth flow; rarely need to manage plugins manually
- •Multiple accounts per plugin (e.g., many Gmail/Slack/Calendar)
- •Useful for entrepreneurs running multiple businesses
- •She notes other platforms are starting to adopt multi-connector ideas
- 7:03 – 8:03
The big limitation: Grok Bot is still single-player (no true group chat)
Claire calls out what OpenClaw/Eve-style systems still do better: multiplayer. While Grok Bot now supports sharing bot templates, it’s not the same as collaborating with one bot in a shared conversation with other people.
- •No “put the bot in a group chat” experience yet
- •Template sharing exists but isn’t true multiplayer
- •Multiplayer becomes important once you rely on agents daily
- •She explicitly requests this feature from the product team
- •Sets expectations: Grok Bot excels elsewhere despite this gap
- 8:03 – 10:05
Chief: her chief-of-staff bot for email/calendar/Slack triage
Claire introduces Chief as the replacement for her beloved OpenClaw “Polly.” Chief runs frequent sweeps across multiple inboxes and calendars, surfaces only high-signal items, and provides briefings and alerts to keep her aligned.
- •Hourly sweeps of inbox, calendar, and Slack across many accounts
- •Only pings Claire when human attention is needed; archives/ignores low-signal items
- •Morning briefing + weekend plan-ahead routines
- •Example alerting: noticing website traffic spikes and advising timing
- •Chief effectively replaces her prior OpenClaw assistant
- 10:05 – 12:38
Tuning Chief: voice cloning, clipped writing, and routine explicitness
Claire shares practical tuning lessons: Grok Bot can write too clipped/telegraphic, so she trains Chief on her email voice. She also notes Grok Bot requires more explicit routine setup than OpenClaw’s “heartbeat”-style proactivity.
- •Grok model output can be overly brief and awkwardly compressed
- •She trains Chief by extracting patterns from her real emails (tone, sign-offs, formatting)
- •Still prefers limited auto-writing for low-stakes messages
- •OpenClaw felt more naturally proactive; Grok needs clearer routine schedules
- •Key takeaway: manage routines explicitly in settings to get reliability
- 12:38 – 14:08
TradBot: a household/family agent built to get you off the computer
TradBot supports Claire’s busy home life (three kids, sports, school logistics) with the goal of reducing screen time. It produces a family “kitchen table newspaper,” anticipates pickup needs, and coordinates weekend planning.
- •Design principle: a home agent should reduce computer time, not increase it
- •Creates a weekday digest as a printable kitchen-table newsletter
- •Includes schedule reminders, weather, kid updates, and parent task lists
- •Helps catch overlooked homework/obligations early
- •Runs just-in-time prompts for school pickup logistics
- 14:08 – 18:40
TradBot workflow details: printing, Kindle reading, and weekend coordination
Claire drills into the artifacts and timing that make TradBot work: printing a newspaper-style summary, optionally emailing PDFs to a Kindle, and running Friday previews to coordinate complex weekend sports schedules across the city.
- •Printed newsletter makes info accessible to the whole family
- •Kid-friendly news topics create dinner-table conversation starters
- •Email-to-Kindle workflow for reading summaries without phone/laptop
- •2:30pm pickup prompt: who’s picking up, what to bring, conflicts
- •Friday preview: weekend schedule, carpools, and gear reminders
- 18:40 – 19:41
Sponsor break: Hyperagent for always-on proactive agents
A sponsor segment describes Hyperagent as a platform for deploying agents that run continuously across inboxes, CRMs, and messaging tools. The pitch emphasizes proactivity, playbooks, guardrails, and quick setup without local infrastructure.
- •Deploy agents where work happens (Slack/Telegram/email)
- •Examples: inbox replies, competitor monitoring, Salesforce deal rescue
- •Proactive agents vs. prompt-waiting chatbots
- •No local setup/VPS; focus on skills/tools/guardrails
- •Offer: $100 free inference for listeners
- 19:41 – 22:12
LGTM (Look Good To Me): the PR Closer to keep engineering flow moving
Claire introduces a bot that manages PR hygiene and throughput, especially important for SOC 2 practices. LGTM reviews the PR queue daily, drives follow-ups, and can trigger Cursor Cloud coding jobs for tasks like rebasing or fixes.
- •Daily sweep of PRs: what to merge/close/rebase/respond to
- •Acts like an “annoying EM” to keep the PR pipeline healthy
- •Supports SOC 2 posture by ensuring review discipline
- •Can kick off Cursor Cloud agents for code tasks (e.g., rebases)
- •Result: cleared a backlog of ~50 PRs and improved queue health
- 22:12 – 23:43
Lockdown: SOC 2 control monitoring and vulnerability triage
Lockdown logs into the SOC 2 control dashboard daily, ensures monitors are green, and escalates anything needing attention. For code-related findings (like vulnerabilities), it can triage and generate a PR, then route approval to Claire.
- •Automates daily checks of SOC 2 monitors and SLAs
- •Logs into web dashboards via the bot’s computer/VM
- •Handles both code controls (vulns) and operational controls (audits/offboarding)
- •Creates PRs to address issues and requests human review
- •Frames AI as “putting us to work” on compliance more effectively
- 23:43 – 27:14
Holly Helpdesk: upgraded customer support with approval-gated actions
Claire describes migrating ChatPRD’s support agent to Grok Bot and seeing noticeable quality improvements—people even ask to leave reviews. Holly sweeps Intercom and email, can trigger fixes via Cursor Cloud, and uses an explicit approval flow for sensitive actions like refunds.
- •Hourly sweep of Intercom + support email using a support playbook
- •Can fix bugs by launching Cursor Cloud agents when appropriate
- •Refunds use an approval-gated ‘agent action’ flow (button + Stripe action)
- •Weekly retro: update docs/playbook based on recurring tickets
- •Also helps manage community Slack by drafting prompts and replies
- 27:14 – 30:16
Personal bots: Penny Pincher (subscription audits) and ShopZilla (shopping research)
Claire shares fun personal workflows: Penny Pincher scans receipts for cancellations, negotiations, and renewal reminders, while ShopZilla handles tedious product research like finding furniture that matches exact dimensions. These bots demonstrate practical “life admin” automation beyond work.
- •Identifies subscriptions to cancel and reminders for upcoming trials/renewals
- •Suggests negotiating big renewals (e.g., home insurance) with market comparisons
- •Flags unused services (e.g., Audible credits) and suggests pausing
- •Can research high-consideration purchases (e.g., a birth-year Rolex)
- •ShopZilla finds products matching constraints (e.g., bookcase dimensions)
- 30:16 – 32:48
Sylvie Style: building a wardrobe plan from Pinterest + weekly sale scouting
Sylvie Style is Claire’s favorite personal bot: it learns her preferences from a public Pinterest board, studies stylist advice, and turns that into a prioritized, shoppable wardrobe plan. It then runs a Monday routine to find sale items that match her style and sends links.
- •Ingests a Pinterest board to infer a coherent “visual language” for style
- •Researches how professional stylists build wardrobes
- •Outputs a prioritized shopping plan (basics, elevated workwear, accessories)
- •Monday 8am routine: scans retailer sales for matching items
- •Sends a curated list of sale picks with direct shopping links
- 32:48 – 34:19
Migrating (and retiring) OpenClaws: Lifeguard export + Grok Bot ‘brain transplant’
Claire explains how she moved from OpenClaw to Grok Bot using a ‘rescue bot’ called The Lifeguard. It exports a secrets-free package of agent identities, cron schedules, and rules so she can upload them into Grok Bot—then cleanly remove the old agents.
- •Uses The Lifeguard bot to export agent setups (identities, crons, rules)
- •Creates a secrets-free zip for migration
- •Uploads subfolders to Grok Bot to recreate behaviors quickly
- •Then removes OpenClaw automation (crons/gateway configs) to ‘kill’ agents safely
- •Positions this as a pragmatic path off hard-to-maintain systems
- 34:19 – 36:16
Final take: UX and reliability beat raw power; invite for viewer use cases
Claire concludes that OpenClaw set an important bar (especially multiplayer), but maintenance overhead was too high even for a technical user. Grok Bot’s simpler UX and “works out of the box” reliability wins for her day-to-day, and she invites viewers to share their own bot ideas.
- •OpenClaw: powerful, innovative multiplayer, great use cases—but fragile to maintain
- •Claire’s technical setup still wasn’t enough to make maintenance painless
- •Grok Bot: simpler UX, strong default experience, “90% coworker agent” value
- •She’s running ~30 bots at any given time
- •Requests audience feedback and ideas for future builds
