HomeVideos

7 Grok Bot agents I use every day

Now Playing

7 Grok Bot agents I use every day

Transcript

931 segments

0:00

Today, I'm going to talk about what

0:01

everybody else is talking about these

0:03

days, Grokbot. Yes, I did an episode

0:06

lovingly entitled Grok Grok Grok, where

0:09

I gave you my high-level overview of

0:12

Grokbot, the new Grok models, and

0:14

Origin, Cursors new GitHub replacement.

0:17

But, today I'm going to talk

0:18

specifically about Grokbot. After

0:21

several weeks of working with Grokbot, I

0:23

have to tell you the truth. I have

0:25

killed all my open claws and moved

0:27

almost entirely

0:30

to Grokbot. In this episode, I'm just

0:32

going to show you a couple very

0:34

practical use cases of Grokbot, why I

0:37

think it's better right now than some of

0:40

the open claws I was running, what is

0:42

still missing, and what I can't wait for

0:45

you to try. This is going to be a

0:46

totally hands-on mini episode with bot

0:49

ideas you can steal. We will put the bot

0:51

links and the bot templates in the show

0:53

notes, and I hope this is a really

0:55

specific way to get you up and running

0:57

in your life a little bit better with

0:59

Grokbot. This episode is brought to you

1:02

by WorkOS. AI has already changed how we

1:05

work. Tools are helping teams write

1:07

better code, analyze customer data, and

1:09

even handle support tickets

1:11

automatically. But, there's a catch.

1:13

These tools only work well when they

1:15

have deep access to company systems.

1:18

Your co-pilot needs to see your entire

1:20

code base. Your chatbot needs to search

1:22

across internal docs. And for enterprise

1:24

buyers, that raises serious security

1:27

concerns.

1:28

That's why these apps face intense IT

1:30

scrutiny from day one. To pass, they

1:33

need secure authentication, access

1:35

controls, audit logs, the whole suite of

1:37

enterprise features.

1:39

Building all that from scratch, it's a

1:41

massive lift. That's where WorkOS comes

1:43

in.

1:45

WorkOS gives you drop-in APIs for

1:47

enterprise features, so your app can

1:49

become enterprise-ready and scale

1:51

upmarket faster. Think of it like Stripe

1:53

for enterprise features.

1:55

OpenAI, Perplexity, and Cursor are

1:58

already using Work OS to move faster and

2:00

meet enterprise demands.

2:02

Join them and hundreds of other industry

2:05

leaders at workos.com.

2:08

Start building today.

2:10

Okay, quickly before we get into use

2:11

cases, let's just remind you of exactly

2:13

what Grok Bot is. Grok Bot is the new

2:16

multi-agent platform from the SpaceX AI

2:19

team, which is really positioned as AI

2:22

teammates you can give real work. And

2:24

this has been my experience. You all

2:27

know I love multiple open claws. I love

2:30

to name my agents specific names and

2:32

give them specific jobs. And Grok Bot

2:34

has a very similar feel where you can

2:37

define agents, give them jobs, and

2:39

assign them work to do.

2:41

Looking at Grok Bot, there's just a

2:43

couple things that I think you need to

2:44

know. First of all, Grok Bot is very

2:47

simple. It looks a little bit like

2:49

iMessage, a little bit like terminal

2:51

with just a couple affordances that make

2:53

it nice to write code and have computer

2:56

use, but there are a couple key

2:58

primitives of Grok Bot I think is

3:00

important to know as you get started.

3:02

First are

3:04

the Groks, the bots. I don't know what

3:05

we call them. I'm going to call them

3:07

bots. The bots. As you can see on the

3:10

left hand of my screen, you can define

3:13

individual agents. To create an agent,

3:16

you just click this plus sign, and then

3:18

you can either search your existing

3:20

bots. Look, I was correct, they're

3:21

called bots, capital B. Or you can

3:23

create a new bot. Fun fact, if you

3:26

create a new bot with a specific name,

3:28

it will infer from that bot's name what

3:31

its job is. So, if you, like me, created

3:33

one of your first bots named Proddy

3:35

McProd, it's going to understand that

3:38

it's probably going to work with you on

3:40

products.

3:41

So, you can see here I've named a couple

3:43

bots, but these are sort of the threads,

3:46

the agents, the individual defined scope

3:50

and context that you're going to be

3:52

working with. And just general

3:54

recommendation, I really like again this

3:57

sort of like bot as colleague, bot as

3:59

coworker, bot per job framework. So,

4:01

this is why it works really well with my

4:03

mindset. is I like to hire my bots like

4:07

I hire people for specific jobs and with

4:09

specific names and that is the kind of

4:11

like fundamental way that Grok bot

4:13

works. The second thing that you need to

4:15

know about Grok bot are plugins. So,

4:17

plugins are combinations of skills and

4:20

connectors that allow you to interact

4:22

with different parts of your stack. Um

4:25

the most common integrations and plugins

4:28

that you're going to use are going to be

4:29

these productivity ones. So, you can see

4:30

the ones I've added at the top. Your

4:32

Gmail, calendar, Google Drive. I have

4:35

Granola, GitHub, Sentry, Chat PRD. I

4:38

have a bunch of them.

4:40

This is where I fell in love with Grok

4:42

bot. The connector experience is really

4:45

good. One, because you'll rarely go into

4:48

this screen. You will rarely I'm showing

4:49

you this screen because I'm trying to

4:51

give you the world tour of Grok bot, but

4:52

you'll rarely have to actually

4:54

explicitly go into the Grok bot screen

4:56

and set up plugins. Often your bot will

4:59

just prompt you to connect the plugins

5:01

that it needs and then you go through a

5:03

really easy O off flow and then you're

5:05

off to the races. The other thing I love

5:08

about Grok bot and I called this out in

5:10

my earlier

5:11

review is that you can connect multiple

5:15

accounts per plugin. And again,

5:18

thank you SpaceX AI team for this killer

5:21

feature um where I have

5:24

six email addresses. I actually have

5:26

more than this. I probably have 10. They

5:27

can all be connected. My Grok bot can

5:29

switch and move between all of them.

5:32

Super simple. In case you missed it,

5:35

OpenAI um released multi-Gmail,

5:39

calendar, and drive, I think. Don't

5:41

quote me. Um connectors for Codex and

5:44

work. So, I I see some of these other

5:46

harnesses kind of like picking up on how

5:48

important this is, but especially as an

5:50

entrepreneur that runs multiple

5:52

businesses, the ability to connect

5:53

multiple slots, multiple linears,

5:55

multiple Gmails, all those sorts of

5:57

things have been super helpful. And it's

5:59

just really elegant the way that these

6:01

connectors are done. The final thing

6:04

that you need to know is that Grok Bot

6:07

has a computer and it has routines. So,

6:10

every Grok Bot has a virtual machine

6:12

which it can do work. This means you can

6:15

log into websites, it means you can

6:17

install things, you can run code

6:19

generation, you can run

6:22

repos locally. So, it does have its own

6:25

computer that you can actually click

6:27

into and yourself run.

6:30

This runs in the cloud. So, if you

6:32

switch like me to Grok Bot on your

6:34

phone,

6:35

you'll have continuity of tasks, which

6:38

is really nice. And then it has

6:41

routines, which is just scheduled times

6:43

it does stuff for you. So, when I talk

6:46

to you about how I design my Grok Bots,

6:48

I'm going to show you how I define them,

6:51

what routines I give them, how I work

6:54

with them, and then how we use plugin

6:56

and computer to get jobs done. Now, the

6:59

caveat that I want everybody to hear,

7:01

including the bot product team, is these

7:04

are still single player. This is the one

7:06

thing where an Eve agent, which I've

7:07

done an episode on, or an Open Claw,

7:10

which I've done many episodes on, really

7:12

still wins is you can't put a Grok Bot

7:14

in the group chat. And I want a Grok Bot

7:17

in the group chat. Once you get addicted

7:20

to the bot, you want to share this with

7:22

other people. They have released a share

7:24

a bot template feature in the past week

7:28

where you can share your bot with your

7:29

team and they can have the same

7:30

instructions and routines and stuff like

7:32

that, but it's just not the same as a

7:34

single bot between two folks chatting

7:37

with it. I really look forward to that

7:39

feature from the team whenever they get

7:41

it live.

7:42

Okay, so that is that was the

7:43

fundamentals of GrokBot. Let's go into a

7:46

couple specific ways I use it. I'm going

7:49

to cover kind of a general purpose chief

7:51

of staff bot. I'm going to look at some

7:53

of my business bots and I'm going to end

7:55

with some of the more fun personal bots

7:58

that I have had a really delightful

8:00

experience with.

8:02

So first, let's talk with Chief. Um

8:05

Chief, the chief of staff bot, you can

8:07

see here Chief has got the salute. Um it

8:10

doesn't even actually have a name. It's

8:12

just named this emoji. But Chief is just

8:16

my general purpose keep my life aligned,

8:19

go through my emails, go through my

8:21

calendar, make sure I'm on top of stuff

8:25

bot.

8:26

Chief is who has killed Polly. If you

8:30

are new here, Polly is my beloved

8:33

OpenClaw. She has deleted my calendar.

8:36

She has responded to emails. She has

8:38

texted people that I'm late. Polly is

8:40

the bar for a gentle co-worker and I

8:43

regret to inform you

8:45

I have murdered murdered Polly. In fact,

8:48

I had Polly kind of murder herself. It

8:50

was very meta. I will show you quickly

8:52

how to migrate from OpenClaw to um

8:55

GrokBot if you want.

8:57

But Chief is my now chief of staff and

8:59

doing quite a good job. So, what do I

9:02

have Chief do? Well, I have Chief do

9:06

basic chief of staff stuff. So, on an

9:08

hourly basis, Chief sweeps my inbox, my

9:11

calendar, and my Slack. And again, I

9:13

have like six inboxes, six Slacks, many

9:17

calendars, many meetings. And it's only

9:20

going to ping me when it needs me. So,

9:24

it's supposed to clear out emails I

9:26

don't need to look at. It's supposed to

9:29

just leave things unread that I don't

9:31

need to worry about, but it's going to

9:33

ping me and be very high signal about

9:35

what I need to pay attention to. And I

9:39

run this hourly, sort of like business

9:41

hours, which for me are like 6:00 a.m.

9:43

to 9:00 p.m. Um, 7 days a week. I don't

9:46

really run it overnight. I I don't want

9:48

people to get used to me responding to

9:49

them overnight, but we'll see.

9:52

And it's going to archive work that

9:53

stays silent. The other thing that I've

9:55

trained Chief is how to talk like me. I

9:58

will say the Grok model sniffs of

10:03

Claude, um, in that it's just really a

10:05

terrible writer. And one of my

10:08

complaints about Grok bot is that it has

10:11

been tuned clearly to be brief and token

10:14

efficient.

10:16

And that means sometimes it just says

10:18

crazy stuff. It tries to compress

10:20

compress a sentence that would make

10:22

sense if you wrote out the sentence into

10:25

like three or four words, and it just

10:27

doesn't have a very natural voice. It

10:29

doesn't have sort of inscrutable Claude

10:32

slop in that it's too long or verbose.

10:34

It's actually the opposite. It's too

10:35

clipped and short, and it's hard to

10:37

understand what it means. So, one of the

10:39

first things I did is said, "Go through

10:41

my emails and really pull out my voice."

10:43

And I will say I've tried this a couple

10:45

times with a couple models, a couple

10:46

harnesses. I've been really happy with

10:48

the prompt that Grok bot put together.

10:50

Again,

10:52

I have short short warm funny emails,

10:55

never an M dash. I usually open emails

10:57

with hey. I usually sign them off with a

11:00

little C. I have scheduling links, all

11:03

that kind of stuff. And it's done a

11:05

pretty good job. I still don't love it

11:07

writing on my behalf, but for low stakes

11:09

stuff, it definitely works.

11:11

So, I've trained

11:12

Grok bot, um, Chief on my voice, and

11:15

then it does that hourly sweep. It does

11:17

a morning briefing, and it does a

11:19

weekend plan ahead, so I know what's

11:21

happening in my work life in the weekend

11:24

ahead. And then it can go back and forth

11:27

with me on alerts that it thinks I need

11:29

to know about. For example, right here

11:32

you can see I got

11:34

um some data transfer edge request

11:36

spikes on the website. I think that's

11:38

because we released the article today.

11:41

I think that's because we released a

11:42

podcast today and a lot of people come

11:44

visit the website after we release a

11:46

podcast. But it said it's worth looking

11:49

at, but you can wait till after your

11:51

meeting. I've been really pleased with

11:52

how Chief works. That being said, I have

11:54

had to tune it and what I've had to tune

11:56

in terms of Grokbot versus OpenClaw that

12:00

I think was a little bit better is

12:01

OpenClaw had that heartbeat function.

12:04

So, if I said, you know, like every 5

12:05

minutes or 30 minutes or whatever, I

12:07

just need you to be more proactive.

12:09

Out of the box, I felt like OpenClaw was

12:11

a lot more proactive. It was hard to

12:14

maintain and hard to get to that

12:15

productivity, but it was sort of like

12:17

more natural. Here, I really had to tell

12:21

Chief a couple times, like put this on a

12:23

schedule. Why aren't you doing this? I

12:25

asked you to watch my emails. And so, I

12:28

would say being a little bit more

12:29

explicit about when you want it to run

12:31

routines, you kind of have to manage

12:33

that. It's easy to see because it's set

12:35

up in the settings, but it is something

12:37

that I noticed that you have to do

12:39

pretty specifically. Okay, so that's

12:42

Chief, my business chief of staff, but

12:44

then of course, we have

12:47

Tradbot. And Tradbot

12:50

is my family agent. Now, I have a very

12:53

complicated business life. I have an

12:55

even more complicated home life. I've

12:58

got three kids, three boys. I've got a

13:01

baby. We're in apparently 80 different

13:04

sports this year. We're teaching

13:06

pre-algebra. We're teaching people to

13:08

read. We're teaching babies to sleep.

13:10

We're doing all the things. I have a

13:13

delightful husband upstairs who takes a

13:15

lot of our family operations off my

13:18

plate. Very, very appreciative of that.

13:21

But we need help and the help we get is

13:23

an agent. And

13:24

this is what I have to say generally

13:26

about parenting {slash} agents for the

13:28

home is the ideal thing for an agent for

13:31

a home is to get you off your computer

13:34

and with your family. So, that is always

13:36

my goal with a trad agent, a household

13:40

agent, is I actually don't want to spend

13:42

a lot of time with the agent. That's a

13:44

failure state. I want the the agent to

13:47

make it really easy for me to spend time

13:50

with my kids

13:51

off the computer. So, I'll talk a little

13:53

bit about what I do with these agents.

13:55

It may sound familiar to you if you've

13:56

listened to my podcast about Open Claw

13:59

about 6 months ago.

14:01

So, um of course we do a weekday family

14:03

digest, but more specifically we do a

14:05

kitchen table newspaper

14:08

that outlines highlights about our day,

14:11

the kids' days, important reminders, but

14:14

is printed off this beautiful um HP

14:16

printer I have over here and gets set on

14:20

the table next to our kids so they can

14:22

read about it, we can all read about it.

14:24

And of course I don't want to dox my

14:25

beautiful children, but I will show you

14:28

a sort of anonymized version of what

14:30

this um newsletter looks like.

14:33

So, it's got a title across the top.

14:35

It's got um a schedule, so who has our

14:38

winter spelling test coming up. It's

14:40

Labor Day, so we're not going to have

14:41

school on Monday. Um we live in San

14:44

Francisco, so it will show the weather

14:46

which is always between 55 and 62 at

14:49

least in um August, and so it'll tell

14:53

tell us what we need to wear, always

14:54

wear a light jacket. I can make this a

14:56

fixed script and it would probably be

14:57

the same. And then it's saying, "Okay,

15:00

school's going to be misty and foggy."

15:02

My kids go to school kind of um in the

15:04

in the fog in the woods, so this is good

15:06

to know. And then for each kid it kind

15:09

of says like, "What do we know from the

15:10

emails from school? What do we know

15:12

about your homework? What's happening up

15:14

this weekend?" And then mom and dad

15:16

tasks, and I will tell you the straight

15:18

truth. I think I am I'm a tiger mom. I'm

15:21

on top of my kids' homework. And this

15:24

morning I read this I read this

15:25

newsletter, not this specific example. I

15:28

read this newsletter, and so it said,

15:30

"Don't forget middle kid has math

15:32

homework due on Wednesday."

15:34

And in my mind I was like, "I haven't

15:36

seen

15:37

math homework." It was math homework was

15:39

not in my head. And you know, he's he's

15:41

little. I want to make sure that I'm on

15:43

top of things, and it was just one of

15:44

those like very nice quality of life

15:46

things. Wasn't on a computer, wasn't an

15:49

email, was printed out on my kitchen

15:52

desk, and allowed me to kind of know

15:55

what was going on and not forget all the

15:57

things that we seemingly are juggling.

15:59

And then my favorite part on the right,

16:01

it has kid-friendly news topics. Um the

16:05

Pokémon World Champions were in San

16:07

Francisco, and so

16:09

and talks about that event, which was

16:11

here. I was really excited about this.

16:13

This came up in a couple of our

16:14

newsletters, but the NASA telescope

16:16

blasted off this week. So just like a

16:18

really cool fun things, fun facts for us

16:21

to talk about um at the kitchen table in

16:24

the morning when we're having breakfast

16:25

or doing our homework that I think is

16:26

really fun. Um I do have to steal a kind

16:30

of like little minor workflow from my

16:32

husband, which is you can use an email

16:35

to Kindle address to email this kind of

16:38

newsletter to your Kindle and read it. I

16:41

think it's such a genius like I idea

16:43

again to get off

16:45

a laptop, get away from your phone, but

16:47

still read the news that you care about

16:50

or have a summary. A Kindle can be a

16:52

nice thing, and you can email any PDF to

16:54

your email with Kindle address. And so

16:56

that can be another way if you make this

16:57

not for your kids, but for yourself,

16:59

that you can ingest this kind of

17:01

information.

17:03

Okay, so um morning digest kitchen table

17:06

newsletter, and then just making me

17:09

aware I am supposed to be. So this uh

17:13

kicks off at 2:30 every afternoon before

17:16

school pick up, and it just asks,

17:19

"Claire, EJ, or someone else, who is

17:21

picking up the kids? What do we need to

17:23

bring? Do we have the jujitsu gear? Do

17:26

we have the soccer snacks? Um do we have

17:29

conflicts where piano's over here and

17:31

baseball's over there? Do we have the

17:32

things?" And again, I hope that parents

17:35

can or maybe I don't hope that they can

17:36

empathize with with this, but we are

17:39

just every day living our um best

17:42

extracurricular life.

17:44

Sometimes I have meetings, sometimes my

17:46

husband has meetings, and just knowing

17:47

getting us really coordinated just in

17:49

time to head out the door and get our

17:50

kids to the right things when it seems

17:52

like every week those things might

17:54

change schedule just a little bit has

17:55

been

17:56

super helpful. And then the mega version

17:59

of that, which is a Friday preview of

18:01

what's ahead on Saturday and Sunday.

18:03

Again, we hate ourselves, so we've

18:05

signed our kids up for three sports

18:07

across the entire city, and sometimes

18:10

one kid needs to be at the marina and

18:12

another kid needs to be in Woodside, and

18:15

then we have a baby.

18:16

And we just need to know who needs to go

18:17

where.

18:18

Just a summary of like, "Here's what's

18:20

happening this weekend. Here's when, you

18:23

know, you have a bye week. Here's what

18:24

you need. Here's when you're carpooling

18:26

someone." has been really, really

18:28

helpful and keeps us in line. I have

18:30

released Chad Bot as a template if you

18:32

want to borrow it, just go to my Twitter

18:34

or look in the show notes and I will

18:35

share you all these templates, but

18:37

super, super useful um bot. I

18:41

love them. This episode is brought to

18:43

you by Hyper Agent, the platform for

18:45

deploying always-on agents that actually

18:48

run your business.

18:50

With Hyper Agent, you build agents in

18:51

the cloud and deploy them where your

18:53

work already happens, like Slack,

18:55

Telegram, or email. An agent will scan

18:58

your inbox and draft replies to vendor

19:00

follow-ups. Another monitors competitors

19:02

and spins up rich ad kits and landing

19:05

pages. A third notices a deal going cold

19:08

in Salesforce and writes the save email

19:10

with full account context. These aren't

19:13

chatbots waiting for a perfect prompt.

19:15

They're proactive, learning your

19:16

preferences, retaining your playbooks,

19:18

and getting better with every run. One

19:21

user built four agents to run an

19:23

outbound sales pipeline. Prospecting,

19:25

outreach, follow-ups, CRM updates, all

19:28

in a single afternoon. No local setup,

19:31

no VPS bills, no fragile permissions on

19:34

your laptop, just powerful agents with

19:36

full control over skills, tools, and

19:38

guardrails. How I AI listeners get $100

19:43

in free inference to start building.

19:45

Claim yours at hyperagent.com/howiai.

19:51

Okay, moving to some engineering use

19:53

cases that I think are quite useful. I

19:55

have two that I will show you both

19:57

around code review and keeping code

20:00

moving in a healthy way. One is look

20:03

good to me, the PR closer. I don't know

20:05

about you all, but I am just opening PR

20:08

after PR after a PR after PR. And we

20:12

have to as part of our SOC 2 compliance

20:15

posture and controls, we do review the

20:17

PRs. We review them both agentically and

20:20

as humans. And like

20:21

PRs are stacking up a little bit because

20:23

it's way more fun to issue a PR than

20:25

review a PR. So, if you have empathy

20:27

that your job has become reviewing PRs

20:29

and you don't like it, I have empathy,

20:31

too. So, I created look good to me, the

20:34

PR closer, which every day goes through

20:38

my PRs and sees what needs to be merged,

20:42

what needs to be closed, what needs to

20:44

be rebased,

20:46

um responds to comments, sends Slack,

20:49

and does all the things to basically be

20:51

like an annoying engineer manager to

20:53

keep things moving along in the PR

20:56

pipeline. Right now, I have a kind of

20:59

kooky bananas stacked PR um with a bunch

21:03

of different branches stacked on top of

21:05

it, all of which need to be rebased and

21:07

monitored for bug bot feedback and all

21:08

this stuff, and look good to me. The PR

21:10

closer is managing all that for me as

21:13

well. Um why I want to show this one is

21:16

uh it can Grok bot can kick off cursor

21:19

cloud agent um coding jobs. And so, you

21:24

can see here that these rebases or

21:28

feedback on different PRs can be done by

21:31

cursor. So, I've also allowed it to and

21:33

given it some rules about writing in the

21:36

code base and updating things. And this

21:37

is just a really nice again quality of

21:39

life thing for our engineering team, so

21:41

that we know exactly what needs

21:43

eyeballs, review stamp, and can be

21:46

merged. And I think with um look good to

21:49

me the PR closer on Friday, I like

21:51

really ran through about 50 PRs that had

21:54

just been sitting on basically me to

21:55

review. Some big, some small.

21:58

Um and so, I got through them and our PR

22:00

queue is much much healthier now.

22:03

The other one, um which I think is a

22:06

really great use case that I would love

22:09

to to call out for folks is we have our

22:12

SOC 2 um controls. They're managed in a

22:15

central dashboard, and every day as a

22:18

small team we have to go through and

22:20

make sure that all of our monitors are

22:21

green, that any monitors that are

22:23

alerting are not so outside SLA, that

22:25

we're closing vulnerabilities, all that

22:27

kind of stuff. And so, I have built

22:30

lockdown the SOC 2 control bot. And what

22:33

this SOC 2 control bot does, is it

22:35

actually logs into that dashboard we use

22:37

to manage our SOC 2 controls, looks

22:40

through all the controls, make sure that

22:42

none of them need human attention. Um if

22:45

there are code controls that we need to

22:47

address, for example, a new

22:48

vulnerability identified, it triages

22:50

that vulnerability and issues a PR and

22:52

then asks me to um review and approve

22:55

it. And this is across both code

22:57

controls and personnels or of

22:59

operational controls, things like have

23:01

we done an account audit? Um and so it's

23:05

just a very nice to have someone, quote

23:08

unquote, an agent monitoring this

23:11

company process for us and making it

23:14

really easy to interact with the

23:16

responsibilities that we have um to our

23:20

clients. And so

23:22

I sometimes say this, we can put AI to

23:24

work or we can have AI put us to work in

23:27

a more effective way. And I really think

23:30

that having AI put me to work against

23:33

our SOC 2 controls and our um compliance

23:37

and security position is really nice.

23:39

It's just like having an intern who taps

23:41

me on the shoulder every now and then

23:42

and says, "Don't forget tomorrow is

23:44

account review today." Or "Don't forget

23:46

we're going to offboard um contractor B

23:48

because they're they're no longer with

23:49

the company and we need to go through

23:51

their offboarding list." Again, we still

23:53

drive these controls, but um having this

23:56

bot to manage that is very helpful. And

24:00

then finally, I do have to say something

24:03

that has been really impressive from a

24:05

quality perspective is migrating over

24:08

Holly Help Desk, our help desk um

24:11

customer support agent at ChatPurity,

24:13

over to GrokBot. Um this again used to

24:16

be an open cloud which was pretty good.

24:18

I spent a lot of time giving it tools

24:20

and access and controls and all that

24:22

sort of stuff. Um but now it runs

24:25

through GrokBot. And why I say I've been

24:27

so impressed is in the last week, since

24:29

running Holly Help Help Desk, I have

24:31

gotten multiple people asking to leave

24:33

reviews about our support being so good.

24:36

And I've not heard that before. I mean,

24:38

you all love me, but no one has said

24:40

like, "Let me leave a review for your

24:42

exceptional customer support." But Holly

24:44

Help Desk is crushing. And so I do have

24:47

to say the quality from a customer

24:49

service perspective is really good. So

24:51

what she does is every hour she sweeps

24:53

the Intercom and email inbox. She uses

24:56

our support playbook and tools, and she

24:59

manages um refunds, technical support

25:03

fixes. She's empowered if there is a bug

25:06

to fix it using cloud um Cursor Cloud

25:10

agents. She is empowered to request

25:14

refunds for me, and I really love this

25:17

um request tool use approval flow. You

25:20

can see here every time a refund needs

25:22

to be um processed, it gives me a

25:25

button. It says, "Should I refund or

25:26

not?" I'm going to go ahead and refund

25:28

Ruben's $19. It should push to me a

25:31

Stripe request. That's Stripe request. I

25:33

can open in Stripe. It's called an agent

25:36

action. It allows me to explicitly

25:38

approve that action for the agent to do

25:40

on my behalf in Stripe, and then this

25:43

the refund is issued, and Holly emails

25:46

back. The other thing Holly does is on a

25:48

weekly basis looks back 7 days and sees

25:52

if there are any items we need to add

25:54

either to our support playbook or to our

25:57

docs to make sure that most of our

25:59

support issues can be self-service. This

26:02

is really helpful. And then finally,

26:04

we have a ChatPurity community Slack.

26:07

And I'm really terrible at managing the

26:09

ChatPurity community Slack. And so I

26:12

have given Holly access to manage the

26:14

ChatPurity community Slack, um look in

26:18

support, and reply to those questions,

26:20

and even prompt and manage um community

26:23

engagement inside that Slack. I'll show

26:24

you what it looks like. So you can see

26:26

here, this is a message from Holly Help

26:29

Desk. Um basically been I've been down

26:31

um heads down on some awesome stuff. Um

26:34

and there's some behind-the-scenes work

26:36

going on in ChatPurity, and please drop

26:38

a hello and instructions. And it says

26:40

sent using Cursor because I did not send

26:41

this. Holly sent this. So again, this is

26:44

one of these things that helps me be

26:46

more engaged with my community. I can

26:47

then go and reply to comments, do all

26:50

sorts of stuff, but I don't have to have

26:51

that blank page or that reminder to do

26:54

this work. Um the bot itself can prompt

26:57

me to do it. Okay, finally, I want to

27:00

end with a couple really fun use cases.

27:04

So, these are more personal use cases

27:07

that I have found to be quite

27:09

delightful, and I hope that you will

27:12

find interesting as well.

27:14

One is Penny Pincher. Penny Pincher does

27:16

exactly what you think she does, which

27:18

is she goes into my email and finds

27:21

pennies to pinch. She finds things to

27:24

cancel. And so, I just say, "Go through

27:26

my receipts, find subscriptions I should

27:28

cancel, things I'm paying for that I

27:30

shouldn't do." You And what was really

27:32

interesting about using Penny Pincher,

27:35

the Grok Bot, is that it did find things

27:37

that I should cancel, like a Apple game

27:40

that I was paying for since my kids were

27:41

really little, and now they're big, they

27:43

don't need it anymore. But, it's also

27:45

found things that I can negotiate that

27:47

will renew in the future. So, a good

27:49

example it found was, "Our home

27:51

insurance is going to renew in a couple

27:53

months. Um I want you to I'm going to

27:56

negotiate, find other rates, and go back

27:58

to our agent with market rates against

28:00

the policy that we have, so we can have

28:03

a negotiated rate."

28:05

It also says that a trial for Apple

28:09

Creator Studio is ending in November.

28:12

It's going to remind me ahead of that

28:13

renewal, and I can decide whether or not

28:15

I want to renew it or not. And then it

28:17

noticed I wasn't using Audible, and it

28:20

said, "You know, you've been paying for

28:21

Audible, but you haven't been actually

28:23

redeeming your Audible credits for a

28:24

couple months." And so again, I paused

28:26

that for 90 days because I have

28:27

aspirations to listen to Audible. I'm

28:29

not going to listen to Audible. Um but,

28:31

it's going to remind me in November when

28:33

that pause is ending, and help me decide

28:35

whether or not I want to do it. Another

28:38

thing that I had Penny Pincher do, which

28:40

is not really penny pinching, it is

28:41

really spending dollars, is I wanted a

28:43

birth year Rolex. I just I want a fancy

28:45

watch. I've been wearing this Apple

28:47

Watch for my entire adult life. Um And

28:50

so I said, "Can you go shop for me?" And

28:52

it actually went into the marketplace

28:54

and found the exact size, which is not

28:55

the standard size, and a couple options

28:58

in particular ones that were

29:01

um

29:02

ones that were released in the year that

29:03

I was born, and it gave me a bunch of

29:05

options, and then it told me what the

29:08

market price should be against the

29:10

watches on the market, and then

29:11

identified ones that were open for

29:13

negotiation and started negotiating for

29:16

me. Now,

29:17

I'm going to tell you all who I really

29:19

am. I'm just I'm too low maintenance to

29:22

buy this watch. So, I will be staying

29:23

with my Apple Watch.

29:25

Um if you all want to convince me that a

29:27

grown-up watch is worth it, please do.

29:29

But, I thought it was that was pretty

29:31

interesting to have it go shop for a

29:34

watch and even a negotiate a good price

29:36

for this. I have created separate from

29:38

Penny Pincher a Shopzilla bot that can

29:41

just go shop things for me. I am trying

29:44

to find a 30-in wide, 15-in deep

29:48

bookcase for a corner of my house, and

29:51

you can see here Shopzilla went out on

29:54

their its browser,

29:56

and it found me a bunch of different

29:58

options that I can purchase for or buy

30:02

that all match the dimensions. This is

30:04

something that's a total pain that I

30:06

would never want to do manually, and it

30:08

really did it quite well and found some

30:10

options that I wasn't able to find

30:12

myself. And then the most fun personal

30:15

use case, I will just end today's

30:17

episode

30:18

with my favorite Grok bot, which is

30:21

Sylvie Style. She is going to help me

30:24

buy clothes.

30:25

Um Just had a baby. I'm still nursing

30:28

that baby. I don't like my clothes.

30:30

They're not fun. I'm a podcaster, so I

30:32

really only have to have cool clothes

30:34

from the waist up. But, I have this

30:36

beautiful aspiration board as many of us

30:39

do about like what I would like to wear

30:42

if I were human that left the house. And

30:44

so I have this Pinterest board. And so

30:46

what I gave Sylvie style is I gave her

30:49

that Pinterest board. I made it public

30:51

so she could look at it.

30:53

She logged in

30:54

and she went and combed through and

30:56

created a visual language for herself

30:59

about my style. I asked her to look on

31:02

Reddit for personal stylist and stylist

31:05

how they approach building people's

31:07

wardrobe. So I was like, go find out

31:08

what the experts do.

31:10

And then Sylvie style built me

31:13

my own style board including top

31:18

priority items that I need or should buy

31:22

um including basics,

31:25

sweatpants because I was like, look I'm

31:27

doing school drop off. We're not trying

31:28

to be fancy.

31:30

Some nice up-leveled workwear

31:32

and then um some fun accessories like

31:36

graphic tees and jewelry. And so this is

31:40

really fun because again, this is like

31:42

kind of a thing I like to do on my free

31:43

time which is go shop and find things I

31:46

like, but I really don't have I don't

31:48

know. I just don't have confidence that

31:49

I'm buying the right things, that

31:50

they're going to all go together, that

31:52

it's worth the time. And so being able

31:55

to say like, this is what I like, these

31:57

are the styles I like,

31:58

build me a plan of high quality things

32:01

that I can invest in and then give me a

32:03

shoppable board has been super awesome,

32:06

but the cherry on top is the Monday

32:10

wardrobe shop. So every Monday

32:14

at 8:00 a.m. it goes through and sees

32:17

what retailers are running sales and

32:21

then goes and finds pieces inside those

32:23

sales that match my style and sends me a

32:27

message with links to shops. So I'm

32:30

going to show you what that looks like.

32:31

So it looks like this. Um here's the

32:34

Monday sale haul um given me, you know,

32:37

about 10 things that I can go look at

32:40

that match what I would like to buy that

32:42

are actually on sale, so I'm saving a

32:45

little bit money, um, that I can go

32:47

purchase and maybe build into my

32:49

wardrobe if I would decide to wear

32:51

anything but AI startup hoodies. Okay,

32:55

so that has been my world tour of Grok

32:58

Bot. I'm going to leave you with a final

33:00

tip, which is

33:02

how to execute your open claws. And, um,

33:06

I did this sadly, I did this with no

33:08

joy, but I did do it and I think it's

33:10

worth showing you.

33:12

So, I have a rescue bot because when

33:15

you're running open claw, you have to

33:16

have a rescue bot to save all your other

33:19

open claws because they're a nightmare

33:21

to maintain, but I have this rescue bot.

33:23

His name is the lifeguard. His only job

33:25

is to manage the other bots and

33:28

I said, "Hey, I want you to pull a

33:29

secrets free package of my entire agent

33:31

setup, crons, agent identities, rules,

33:33

etc." Um, I'm going to move everything

33:35

to Eve agents, but then I moved

33:37

everything to, um, Grok bots. I'm really

33:39

struggling with Gog off and it's making

33:41

it impossible to use. So, it went

33:42

through and actually provided me

33:45

a, um, zip file that had all my agent

33:48

identities, all my cron files,

33:51

everything I needed, um, and I was able

33:54

to just literally upload sub folders of

33:57

that to, um, Grok Bot and it kind of

34:00

like did brain transplant. It pulled

34:02

over their personalities, pulled over

34:04

their tasks, pulled over their

34:05

connectors, and I was ready to go. And

34:07

then, when it was time to

34:10

kill the agents,

34:12

I just said,

34:14

"Kill Sam." I migrated him and the

34:17

lifeguard went through and pulled all

34:20

the crons, pulled all the, um, gateway

34:22

configuration, pulled everything out and

34:25

Sam went to sleep. So,

34:27

not to end on the most morbid part of

34:29

the episode, but that is how to migrate

34:33

your beloved open claws

34:35

to a new platform and then

34:39

get rid of them. Now,

34:41

what do I have to say having gone

34:42

through this?

34:43

UX matters. It just does. And I think

34:47

open claws set the standard for what an

34:49

autonomous teammate personal agent can

34:51

do. Multiplayer was genius. Um use cases

34:56

were awesome. But at the end of the day,

35:00

I struggled so much to maintain my open

35:02

claw, and I am a very technical person.

35:06

I have tailscale running. I'm SSH'd into

35:09

my Mac mini. I have a lifeguard claw. I

35:12

have all this stuff, and it was just too

35:14

hard. And so, I think um that SpaceX AI

35:18

team really nailing a simple everything

35:21

works out of the box 90%

35:24

co-worker agent um has really just

35:26

solved a lot of my problems, and I have

35:28

let these agent rip. So, I've showed

35:30

you, you know, a half dozen of them. I

35:32

have probably 30 running at any one

35:34

time. I would love to hear from you

35:37

what your use cases are for Grok Bot, if

35:39

there's any you missed, if there's any

35:41

you want me to build from you. And in

35:43

the meantime, enjoy your bots, and tell

35:46

me what you get done with them. Thanks

35:48

for joining How AI. I'll see you soon.

35:52

Thanks so much for watching. If you

35:53

enjoyed this show, please like and

35:55

subscribe here on YouTube, or even

35:57

better, leave us a comment with your

35:58

thoughts.

35:59

You can also find this podcast on Apple

36:01

Podcasts, Spotify, or your favorite

36:04

podcast app. Please consider leaving us

36:07

a rating and review, which will help

36:08

others find the show.

36:10

You can see all our episodes and learn

36:13

more about the show at how AI pod.com.

36:17

See you next time.

Interactive Summary

The video provides a detailed, hands-on review of Grokbot, a new multi-agent platform. The host shares her transition from using open-source "open claws" to Grokbot, highlighting its simplicity, effective connector experience, and robust routine management. Through several practical use cases—ranging from a professional 'Chief of Staff' for business management and engineering PR review bots to personal bots for household organization and fashion—the host demonstrates how Grokbot can act as an effective autonomous co-worker. Although she notes the current limitation of single-player usage, she concludes that the platform's superior user experience and ease of maintenance make it a compelling choice for managing complex daily tasks.

Suggested questions

4 ready-made prompts