1#### Start1#### Start
2 2
3# Use cases3# Use Cases
4 4
5The best Grok Bot roles own a repeatable outcome, not a loose category of5The best Bot roles own a repeatable outcome, not a loose category of questions.
6questions. Start with read-and-prepare work, review the result, then add6Start each role with read-and-prepare work, review the result, and only then add
7approved actions or a routine.7approved actions or a [routine](/grok-bot/skills-routines-and-automations).
8 8
9Every starter prompt below ends at a review point on purpose. Keep sending,9Every starter prompt below ends at a review point on purpose. Keep sending,
10publishing, purchasing, and production changes behind10publishing, purchasing, and production changes behind
11[approval](/grok-bot/approvals-security-and-privacy#review-an-action).11[approval](/grok-bot/security#approvals-and-auto-review).
12 12
13## Sales Outbound13## Sales outbound
14 14
15**Owns:** account research, contact prioritization, and review-ready outreach.15Owns account research, contact prioritization, and review-ready outreach.
16 16Connect your CRM, product-intent sources, and email.
17**Connect:** customer relationship management (CRM), product-intent sources,
18company websites, email, and professional networks as permitted by their terms.
19
20**Start with:**
21 17
22> Research the 25 accounts in this CRM view. Score them against our ideal18> Research the 25 accounts in this CRM view. Score them against our ideal
23> customer profile (ICP) and recent intent, identify up to three relevant19> customer profile and recent intent, identify up to three relevant contacts
24> contacts per account, and draft email and LinkedIn outreach in the style20> per account, and draft email and LinkedIn outreach in the style examples
25> examples attached. Skip anyone already in an active sequence. Return a review21> attached. Skip anyone already in an active sequence. Return a review list;
26> list; do not send or enroll anyone.22> do not send or enroll anyone.
27 23
28After the output is reliable, create a nightly research routine that stops at24Once the output is reliable, schedule a nightly research routine that stops at
29the review list.25the review list.
30 26
31## Talent Scout27## Talent scout
32
33**Owns:** sourcing, candidate research, outreach drafts, and scheduling
34preparation.
35 28
36**Connect:** applicant tracking system (ATS), approved sourcing tools, email,29Owns sourcing, candidate research, and outreach drafts. Connect your applicant
37and calendar.30tracking system, approved sourcing tools, email, and calendar.
38 31
39**Start with:**32> For this role description, find 20 candidates who meet the must-have
40
41> For this role description, find 20 potential candidates who meet the must-have
42> criteria. Exclude anyone already in our ATS, explain the evidence for each33> criteria. Exclude anyone already in our ATS, explain the evidence for each
43> match, and draft personalized outreach in my voice. Do not contact anyone.34> match, and draft personalized outreach in my voice. Do not contact anyone.
44 35
45Add approvals before external outreach, and have the Bot respect candidate36Keep external outreach behind approval, and have the Bot respect candidate
46privacy, regional requirements, and source terms.37privacy and each source's terms.
47
48## Paid Media
49 38
50**Owns:** campaign monitoring and budget recommendations.39## Paid media
51 40
52**Connect:** advertising platforms, analytics, budget spreadsheet, and Slack.41Owns campaign monitoring and budget recommendations. Connect your advertising
53 42platforms, analytics, budget spreadsheet, and Slack.
54**Start with:**
55 43
56> Pull current spend and performance by campaign. Compare it with the monthly44> Pull current spend and performance by campaign. Compare it with the monthly
57> budget and target customer acquisition cost (CAC), then recommend45> budget and target customer acquisition cost, then recommend reallocations
58> reallocations with the supporting46> with the supporting numbers. Draft a Slack update for the growth team. Do not
59> numbers. Draft a Slack update for the growth team. Do not change budgets or47> change budgets or send the message.
60> send the message.
61 48
62Keep campaign changes behind approval even after the analysis becomes a49Keep campaign changes behind approval even after the analysis becomes a
63routine.50routine.
64 51
65## Expense Manager52## Expense manager
66
67**Owns:** weekly expense reconciliation and missing-information follow-up.
68 53
69**Connect:** expense system, email, shared drive, and finance spreadsheets.54Owns weekly expense reconciliation and missing-information follow-up. Connect
55your expense system, email, and finance spreadsheets.
70 56
71**Start with:**57> Build this week's expense summary from the expense system and the attached
72
73> Build this week's expense summary from the expense system and attached
74> policy. Match receipts from the finance inbox, flag missing categories or58> policy. Match receipts from the finance inbox, flag missing categories or
75> policy exceptions, and draft one follow-up per owner. Return the summary and59> policy exceptions, and draft one follow-up per owner. Return the summary and
76> drafts; do not send messages or change reimbursements.60> drafts; do not send messages or change reimbursements.
77 61
78Ask for policy citations on every exception and totals that reconcile back to62Ask for a policy citation on every exception and totals that reconcile back to
79the source.63the source.
80 64
81## Product Performance65## Product performance
82
83**Owns:** targeted performance investigations with evidence.
84 66
85**Connect:** observability, analytics, incident tooling, and source-control67Owns targeted performance investigations with evidence. Connect observability,
86links.68analytics, and incident tooling.
87
88**Start with:**
89 69
90> Investigate the checkout latency increase since yesterday's release. Review70> Investigate the checkout latency increase since yesterday's release. Review
91> dashboards, traces, and flamegraphs; identify the highest-confidence hotspot;71> dashboards and traces, identify the highest-confidence hotspot, and return a
92> and return a short write-up with screenshots and direct links. Separate facts72> short write-up with screenshots and direct links. Separate facts from
93> from hypotheses. Do not change alerts or production settings.73> hypotheses. Do not change alerts or production settings.
94 74
95Use a routine for a recurring health report, not for unsupervised production75Use a routine for a recurring health report, not for unsupervised production
96changes.76changes.
97 77
98## Bug Reproduction78## Bug reproduction
99
100**Owns:** turning reports into reliable reproduction packs.
101 79
102**Connect:** issue tracker, staging environment, browser, and network tools.80Owns turning reports into reliable reproduction packs. Connect your issue
81tracker and a staging environment.
103 82
104**Start with:**83> Read this bug report and reproduce it in staging with a fresh test account.
105
106> Read this bug report and reproduce it in staging using a fresh test account.
107> Return exact steps, expected and actual behavior, screenshots, browser and OS84> Return exact steps, expected and actual behavior, screenshots, browser and OS
108> details, relevant console or network notes, and a minimal test case if85> details, and a minimal test case if possible. Do not use production customer
109> possible. Do not use production customer data.86> data.
110
111Provide approved test credentials through a secure handoff, not chat.
112 87
113## Account Health88Provide test credentials through a
89[secure handoff](/grok-bot/computer-and-apps#take-over-for-a-sensitive-step),
90not chat.
114 91
115**Owns:** risk and expansion signals across a customer portfolio.92## Account health
116 93
117**Connect:** CRM, product usage, support, billing, and customer-success notes.94Owns risk and expansion signals across a customer portfolio. Connect your CRM,
118 95product usage, support, and billing sources.
119**Start with:**
120 96
121> Review the accounts in this portfolio. Combine recent usage, support97> Review the accounts in this portfolio. Combine recent usage, support
122> escalations, renewal timing, and stakeholder activity into a ranked watch98> escalations, renewal timing, and stakeholder activity into a ranked watch
123> list. For each account, include the evidence, why it matters, and a suggested99> list. For each account, include the evidence, why it matters, and a suggested
124> next step. Do not contact customers or edit the CRM.100> next step. Do not contact customers or edit the CRM.
125 101
126Define risk thresholds in the Bot description so the weekly result stays102Define risk thresholds in the Bot's description so the weekly result stays
127consistent.103consistent.
128 104
129## Chief of Staff105## Chief of staff
130
131**Owns:** a source-linked digest of what changed and what needs attention.
132 106
133**Connect:** Slack, email, calendar, meeting notes, and planning documents.107Owns a source-linked digest of what changed and what needs your attention.
134 108Connect Slack, email, calendar, and your planning documents.
135**Start with:**
136 109
137> Review activity since yesterday across my approved channels, inbox, calendar,110> Review activity since yesterday across my approved channels, inbox, calendar,
138> and meeting notes. Return only items that map to the priorities in this111> and meeting notes. Return only items that map to the priorities in this
139> document. For each item, include the source, why it matters, the proposed next112> document. For each item, include the source, why it matters, the proposed next
140> step, and whether I owe a decision. Do not send messages or change meetings.113> step, and whether I owe a decision. Do not send messages or change meetings.
141 114
142Tune the Bot by marking what was useful and what was noise. Then schedule the115Mark what was useful and what was noise, then schedule the digest for a time you
143digest for a time you can review it.116can review it.
144
145## Turn an example into a durable Bot
146 117
147For any role:118## Turn a role into a durable Bot
148 119
1491. Put the job, source systems, output format, and standing boundaries in the1201. Put the job, source systems, output format, and standing boundaries in the
150 Bot description.121 [Bot's description](/grok-bot/bots#edit-a-bot).
1512. Run one real task with a safe scope.1222. Run one real task with a safe scope.
1523. Correct the result until it is reviewable.1233. Correct the result until it's reviewable without you.
1534. Save the successful process as a skill.1244. Save the working process as a
1545. Test it on a second input.125 [skill](/grok-bot/skills-routines-and-automations#save-a-skill), and test it
1556. Create a routine only when retries and failure cases are defined.126 on a second input.
1567. Keep consequential external actions behind approval.1275. Create a
157 128 [routine](/grok-bot/skills-routines-and-automations#create-a-routine) only
158Continue with [Get started](/grok-bot/get-started), then review129 after retries and failure cases are defined.
159[files and results](/grok-bot/files-and-results) and1306. Keep consequential external actions behind approval.
160[approval boundaries](/grok-bot/approvals-security-and-privacy).131
132## Related pages
133
134* [Get started with Grok Bot](/grok-bot/get-started)
135* [Work with Grok Bot](/grok-bot/bots)
136* [Grok Bot for Teams and Enterprise](/grok-bot/teams-and-enterprises)
