Examples - OmniPalm

1

Triage new emails

Ask OmniPalm to scan your inbox and categorize emails by priority.

> Check my inbox and categorize emails by priority. 
  Mark promotional emails as read.

# Scanning inbox...
✓ Found 24 unread emails

📧 High Priority (3):
  - CFO: Q4 Budget Review (needs response)
  - HR: Benefits enrollment deadline tomorrow
  - Client: Contract revision requested

📧 Normal (8):
  - Team updates, meeting notes...

📧 Low/Promotional (13):
  - Marked as read and moved to promotions
                  

2

Draft responses

Generate contextual draft responses for important emails.

> Draft a response to the CFO's budget email. 
  I'll review the budget doc by Friday.

# Drafting response...
✓ Draft created:

Subject: Re: Q4 Budget Review

Hi Sarah,

Thank you for sending over the Q4 budget projections. 
I'll review the document thoroughly and have my feedback 
ready by end of day Friday.

If there are any specific areas you'd like me to focus on, 
please let me know.

Best regards

✓ Draft saved (not sent) - review and send?
                  

3

Set up monitoring

Configure ongoing inbox monitoring for important senders.

> Alert me immediately via Slack when I receive emails 
  from the CEO, CFO, or any @client.com addresses

✓ Email monitor configured:
  - Watching: ceo@company.com, cfo@company.com, *@client.com
  - Alert channel: #my-notifications
  - Status: Active
                  

1

Create ticket from bug report

Convert bug reports into properly formatted Jira tickets.

> Create a Jira ticket: Users report login fails on Safari. 
  Check error logs for details. High priority.

# Searching error logs...
Found: "Auth token parsing error on WebKit browsers"

# Creating issue...
✓ Created issue #247 on meta/omnipalm

Title: Authentication fails on Safari browsers
Labels: bug, high-priority, authentication
Assignee: @auth-team

Body:
## Description
Users are experiencing login failures specifically on Safari...

## Error Details
```
Auth token parsing error on WebKit browsers
Stack trace: TokenParser.js:142
```

## Steps to Reproduce
1. Open Safari 17+
2. Attempt to log in...

## Environment
- Safari 17.2, macOS Sonoma
                  

2

Run tests and deploy

After fixing the bug, run tests and deploy to staging.

> Run the full test suite for the auth module, 
  then deploy to staging if all tests pass

# Running tests...
✓ Unit tests: 89/89 passed
✓ Integration tests: 34/34 passed
✓ E2E tests: 12/12 passed
✓ Safari-specific tests: 8/8 passed

# All tests passed. Deploying to staging...
✓ Build completed (2m 34s)
✓ Deployed to staging.example.com

📊 Deployment Summary:
   Branch: fix/safari-auth-247
   Commit: f8a2b1c
   Tests: 143 passed
   Coverage: 91.2%
   URL: https://staging.example.com
                  

3

Create PR and notify team

Create a pull request with the fix and notify the team.

> Create a PR for the Safari auth fix. 
  Request review from @alice and notify #eng-team on Slack

# Creating pull request...
✓ PR #312 created: "Fix Safari authentication token parsing"

  - Base: main ← fix/safari-auth-247
  - Reviewers: @alice
  - Linked issue: #247
  
# Notifying team...
✓ Posted to #eng-team:
  "🔧 Code review ready: Safari auth fix
   Fixes PALM-247 | Staging deployed | Tests passing
   Review link sent to team"
                  

1

Search and filter papers

Find relevant papers on a specific topic with filters.

> Find papers on "attention mechanisms in transformers" 
  published in 2024, sorted by citations

# Searching arXiv, Semantic Scholar, ACL Anthology...
✓ Found 156 papers matching criteria

📄 Top 5 by citations:

1. "Efficient Long-Range Attention" (Chen et al., 2024)
   Citations: 234 | arXiv:2401.xxxxx
   → Linear attention with O(n) complexity

2. "Multi-Query Attention Revisited" (Kim et al., 2024)
   Citations: 189 | ICML 2024
   → Memory-efficient attention for large models

3. "Sparse Attention Patterns" (Lee et al., 2024)
   Citations: 156 | NeurIPS 2024
   → Learned sparsity patterns for transformers

[Full list saved to ~/research/attention-papers-2024.json]
                  

2

Summarize key papers

Generate summaries of the most relevant papers.

> Summarize the top 3 papers and highlight their 
  main contributions and limitations

# Downloading and analyzing papers...

📝 Summary: Efficient Long-Range Attention

Main Contribution:
- Introduces "Sliding Window + Global" attention pattern
- Achieves O(n) complexity while maintaining accuracy
- 3x faster than standard attention at 16k context

Limitations:
- Requires specialized kernels for efficiency
- Small accuracy drop on long-range reasoning tasks

[Similar summaries for papers 2 and 3...]

✓ Full analysis saved to ~/research/attention-summary.md
✓ BibTeX entries saved to ~/research/refs.bib
                  

3

Create literature review outline

Generate a structured outline for a literature review section.

> Create a literature review outline covering the 
  evolution of attention mechanisms from 2017-2024

✓ Outline generated:

# Attention Mechanisms: A Survey (2017-2024)

## 1. Foundations (2017-2018)
   1.1 Original Transformer Architecture
   1.2 Self-Attention Mechanism
   1.3 Multi-Head Attention
   
## 2. Efficiency Improvements (2019-2021)
   2.1 Sparse Attention (Child et al., 2019)
   2.2 Linear Attention (Katharopoulos et al., 2020)
   2.3 Longformer & BigBird (2020)
   
## 3. Modern Approaches (2022-2024)
   3.1 Flash Attention
   3.2 Multi-Query & Grouped-Query Attention
   3.3 State Space Models as Alternatives

✓ Outline saved to ~/research/lit-review-outline.md
                  

1

Find optimal meeting time

Find availability across multiple participants' calendars.

> Schedule a 1-hour meeting with Alice, Bob, and Carol 
  next week. Prefer mornings.

# Checking calendars for alice@, bob@, carol@...

Available slots next week (mornings preferred):

  ✓ Tuesday 10:00 AM - 11:00 AM (all available)
  ✓ Wednesday 9:00 AM - 10:00 AM (all available)
  ○ Thursday 11:00 AM - 12:00 PM (Bob has soft conflict)

Recommended: Tuesday 10:00 AM

Create meeting for Tuesday 10 AM?
                  

2

Create meeting with agenda

Create the calendar event with a structured agenda.

> Yes, create it. Title: "Q1 Planning Session". 
  Add agenda: 1) Review Q4 results 2) Set Q1 goals 
  3) Assign responsibilities

# Creating meeting...
✓ Meeting created: "Q1 Planning Session"

  📅 Tuesday, Jan 14, 10:00 AM - 11:00 AM
  📍 Zoom (auto-generated link)
  👥 Alice, Bob, Carol (invites sent)
  
  📋 Agenda:
  1. Review Q4 Results (15 min)
  2. Set Q1 Goals (25 min)  
  3. Assign Responsibilities (15 min)
  4. Q&A / Wrap-up (5 min)

✓ Calendar invites sent to all participants
                  

3

Post-meeting summary

Generate and distribute meeting notes after the meeting.

> The meeting ended. Key decisions: 
  - Q1 revenue target: $2M
  - Alice owns marketing, Bob owns engineering
  - Next check-in in 2 weeks
  Send summary to attendees and post in #team-planning

# Generating meeting summary...

✓ Summary created and distributed:

📧 Email sent to: Alice, Bob, Carol
   Subject: Q1 Planning Session - Summary & Action Items

📝 Posted to #team-planning:
   
   **Q1 Planning Session - Summary**
   
   **Decisions:**
   • Q1 Revenue Target: $2M
   • Marketing Lead: Alice
   • Engineering Lead: Bob
   
   **Action Items:**
   • [ ] Alice: Draft marketing plan by Jan 20
   • [ ] Bob: Engineering roadmap by Jan 20
   
   **Next Meeting:** Jan 28, 10:00 AM
                  

Examples & Use Cases

Email Assistant