12 Grok Bot Use Cases That Feel Illegal

Overview

The video showcases practical “personal AI assistant” automations centered on Grockbot (with examples that also apply to similar assistants like Muse and Dot). The creator emphasizes bots that reduce tedious work by connecting to real services—such as email, calendars, APIs, utilities accounts, and shopping/ride platforms.

---

Key Grockbot use cases, products/features, and how they work

1) Money Saver (subscriptions + negotiation + cancellations)

• Scans all emails to find subscriptions.
• Can negotiate with vendors to lower prices.
• Identifies subscriptions that appear unused and can cancel them.
• Example: found $1,000/month in fixed costs across three Gmail accounts.
• Generates “send-ready” drafted negotiation emails (user clicks send).

Additional included features:

• Checks the California unclaimed property site for assets under the user’s name.
• Can submit relevant forms once the bot identifies what’s owed.

Template/prompt availability: The creator says Grockbot templates + prompts will be provided (hosted via Box).

---

2) Tesla Bot (car control via official API)

• Connects to Tesla Model Y using Tesla’s official fleet API.
• Downloads relevant API documentation, obtains an API key, and uses it with the user’s account context.

Supported actions include:

• “Get the car ready”:
  • sets cabin climate/seat heat preconditions
  • pulls destinations from the connected calendar (when available)
• Garage control (with future expectation to pull the car out)
• Lock/unlock and other controls listed in the prompt

User-only actions are protected:

• signing in
• 2FA codes
• pairing a virtual key

---

3) Computer Cleanup Bot (scheduled disk space management)

• Runs weekly (example: Monday).
• Scans for:
  • cached files
  • stale files
  • oversized files

Safety-first behavior:

• It does not immediately delete.
• It lists items and estimates potential savings.
• Example findings: camera clips, dated output, OBS recordings, cache-like items.
• Clearly flags “high-risk deletion” items.
• Applies rules like never delete empty trash or move files without approval.
• Shows full file paths and reasons.

---

4) Daily Brief Bot (work + family variants)

• Creates a morning daily brief by reading:
  • connected calendar
  • emails
  • to-do list
• Prioritizes urgent and actionable items:
  • urgent emails first
  • today’s meetings with context (attendees and notes when available)

Privacy/safety guardrails:

• read-only unless explicitly approved
• stays quiet when nothing urgent is found

---

5) Family Logistics Bot (kids/sports/school automation)

• Monitors emails from schools/teams and summarizes only what matters:
  • dates, deadlines, forms, payments
  • what to bring
  • schedule changes
• Sends a concise daily digest to a family channel (Telegram group mentioned).
• Updates a shared family calendar:
  • adds events
  • invites the other adult (e.g., wife)
  • avoids duplicate events
• Sends “heads up” the day before events (what to wear/bring, carpool, snacks, sign/pay reminders).
• Example: soccer game logistics translated into concrete tasks (saved ~30 minutes).

---

6) Lead Bot (DM triage + sponsor outreach)

• Scans DMs across platforms (Instagram, LinkedIn, X, TikTok, etc.).
• Categorizes messages into:
  • sponsor/brand deal leads
  • “cool projects” from other people
• Filters out spam/vague pitches.
• For legitimate leads:
  • provides summaries + suggested next steps
  • drafts replies for sponsor outreach
• Allows replying from within the assistant (user clicks send).

---

7) Meeting Summary Bot (Fathom transcription + action items)

• Uses Fathom meeting recorder/transcripts.
• After each meeting:
  • fetches transcript via Fathom API
  • summarizes key decisions and open questions
  • extracts action items and adds them to the to-do system
• Can proactively draft follow-up emails when appropriate, but requires approval before sending to others/publicly posting.
• Works with the “spoken-in-meeting” pattern:
  • if someone says “remember to send the proposal,” the bot turns it into an action item.

---

8) To-Do Manager (To-Doist integration)

• Connects Grockbot to the user’s To-Doist account (API/integration).
• Automatically:
  • adds tasks from emails/meetings/messages
  • organizes tasks with titles/projects/due dates
  • marks completion when the user confirms
  • periodically scans for stale/completed tasks

Safety rules:

• never permanently delete tasks without approval
• prefer completing/rescheduling over deleting
• avoid duplicates

---

9) Email Triage Bot (multi-address workflow)

• Scans multiple inboxes (personal, old work, “Forward Future”).
• Auto-archives low-risk emails:
  • delivery notifications
  • receipts
  • invoices
  • promo noise
• For higher-risk emails:
  • walks the user through them one-by-one with summaries
  • drafts replies for approval (including voice-first while driving)
• For contracts/signature requests:
  • flags e-sign documents prominently
  • avoids auto-archiving contract/legal mail

---

10) Utility Bill / Plan Optimization Bot (PG&E example)

• Logs into a utility account (PG&E-style example).
• Reviews about 12 months of usage patterns.
• Compares eligible rate plans and identifies the best one.
  • Example: ~$1,000/year savings
• After the user chooses, it proceeds through switching steps.
• Requires explicit approval; avoids credential sharing beyond user login steps.

---

11) Calendar Assistant (scheduling + conflict detection)

• Connects multiple calendars (work/personal/family) into one assistant context.
• Capabilities:
  • propose scheduling next week based on availability
  • negotiate meeting times with others
  • detect conflicts (e.g., soccer vs movie)
  • prefer rescheduling over declining when possible
  • schedule appointments (dentist/doctor/haircut) in open time slots

Strong safety rules:

• no create/move/decline/book without “okay”
• labels times in user time zone
• reports what changed after edits

---

12) Uber Bot + DoorDash Bot (real-world ordering/transport)

• Supports:
  • ordering a ride to a destination by a time (e.g., SF by 2 p.m.)
  • re-ordering past deliveries and targeting delivery time
• Coordination claim: other agents can schedule rides/deliveries based on the broader context.
• Example: daily brief finds an hour gap and suggests placing a food order to arrive between meetings.

---

Prompts/templates delivery

• The creator repeatedly states:
  • Grockbot templates will be provided for one-click copying.
  • Prompts (including Grockbot equivalents for other assistants) will be compiled into a document hosted on Box (Box link “down below”).

---

Main speakers/sources

• Main speaker/source: The video creator (unnamed in the subtitles) demonstrating Grockbot use cases.
• Referenced/credited external voices: Brian Johnson and Andrew Huberman (mentioned as sources of discovery for Eight Sleep).
• Referenced systems/tools/APIs/services:
  • Tesla official fleet API
  • Fathom (meeting recorder/transcripts)
  • To-Doist
  • Gmail/email accounts
  • California unclaimed property site
  • Telegram
  • Uber
  • DoorDash
  • PG&E-style utility portals