By Jonathan Peslar | Founder, LeadPanther
Word Count: 4,600+ words | Reading Time: ~18 minutes
Part 1: Why Your AI Delegation Sucks
Telling an LLM to "act as my assistant" is lazy and doesn't work.
It is the biggest trap in AI right now.
When you type "act as my assistant," you think you are getting a $150K Chief of Staff.
You are not.
The AI averages every assistant on the internet.
You get a blend of a Fortune 500 Executive Assistant and a college intern.
You get the exact middle. You get the average.
And average creates more manual work for you.
There is a massive divide happening in business right now.
It is the gap between chatbot users and operators.
Chatbot users tinker.
They spend 40 minutes having a conversation with AI to write one email.
That is playing Engineer.
Serious operators don't play Engineer. We play for ROI.
Chatbot users build conversations.
Operators build systems.
If you are using regular Claude, you are still in the chatbot era.
Every time you open a new tab, your AI has amnesia.
You have to spend 20 minutes explaining your brand voice.
By the time the AI is ready to work, you could have done the task yourself.
Claude Dispatch fixes this.
It executes tasks from your phone directly to your desktop.
It has persistent memory. It takes over your mouse.
It uses your apps and runs scheduled tasks while you sleep.
But a new tool isn't enough.
The gap between a basic "chatbot" and a "Chief of Staff" is not a better prompt.
It is Frameworks.
Without a framework, Dispatch will just execute the wrong task, faster.
You need a system for delegation.
I developed the 4-Element Delegation Formula.
I use this exact system to run operations across 4 SaaS products (LeadPanther, GetDeals, GetInterviews, and AI Executive Outbound) as a solo founder.
It is the meta-framework that powers every single workflow in this guide.
1. Context
Claude needs to know your exact business reality.
Don't assume it knows your pipeline or your target market. Define the situation.
2. Constraint
This is the most critical step. Real experts work with limits.
Give it a strict budget. Give it a tight timeline. Limit its scope.
Without constraints, AI hallucinates success.
Force it to operate inside a box.
3. Format
Never ask for "an analysis" or "some ideas."
Demand a specific, structured output.
Ask for a 2-page executive memo. Ask for a .csv file. Ask for an exact email draft.
4. Verification
Do not trust the AI blindly.
Force it to confirm its own accuracy before delivering the final result.
Here is the exact copy-paste formula to initialize any Dispatch task:
CONTEXT:
You are my Chief of Staff. We are a [Niche] company selling to [Target Audience].
Here is the background on the task: [Insert Detail]
CONSTRAINT:
You have a strict limit of [Time/Budget/Scope].
Do not use external web sources. Only use the local files provided.
FORMAT:
Output the final result as a [Spreadsheet/Memo/Draft].
Structure it exactly like this:
- [Heading 1]
- [Heading 2]
VERIFICATION:
Before finalizing, audit your work against the constraints.
If you missed anything, revise before delivering the file.
Without these 4 elements, delegation fails.
With them, Dispatch becomes your highest-paid employee.
Part 2: What Claude Dispatch Actually Is
Drop what you are doing.
Most people are entirely missing the biggest shift in knowledge work since GPT-4.
You are probably still copying and pasting text from a browser window.
You are doing manual labor with a spellchecker.
Claude Dispatch officially kills the "Chatbot" era.
It is not a chat window.
It is a single, persistent conversation thread with Claude.
And it syncs directly between your phone and your computer's hard drive.
Think of it as a remote employee who lives on your desktop but texts you updates.
How It Works Step-By-Step
You no longer sit at your desk to manage AI.
You manage it on the go.
→ You text Claude from your phone while grabbing a coffee.
→ Claude reads your text and figures out what kind of work is needed.
→ It quietly spins up the right session on your computer back home.
→ If it is a complex coding task, it triggers Claude Code.
→ If it is deep knowledge work, it triggers Cowork.
→ It executes the work using your local files, plugins, and desktop apps.
→ When the job is done, it sends a push notification directly to your phone.
It doesn't give you instructions on how to do the work.
It does the actual work.
Key Capabilities
Dispatch is not an iterative update. It is an entirely new architecture.
Here is exactly what makes it capable of replacing human headcount.
Phone-to-Desktop Delegation
You send a text message from the mobile app.
The actual heavy lifting happens on your computer's local CPU.
You are the manager issuing orders. The desktop is the worker executing them.
Computer Use
Dispatch can literally take over your mouse and keyboard.
→ It opens specific apps on your machine.
→ It opens messy SaaS dashboards and clicks buttons.
→ It updates working spreadsheets with live formulas.
→ It uses your web browser to scrape competitor pricing.
Persistent Memory
Regular AI models act like goldfish.
They forget everything the moment you close the tab.
Dispatch remembers your preferences, your active projects, and all past work.
No more re-explaining your business model every single morning.
Scheduled Tasks
This is where true, hands-off automation happens.
You can set up routines that run automatically without being asked.
→ Daily CRM hygiene sweeps at 5 PM.
→ Weekly competitor ad audits on Monday mornings.
→ Monthly financial reconciliations on the 1st of the month.
Direct File Access
Dispatch reads and writes directly to your local folders.
You give it a messy "Downloads" folder.
It organizes the files, extracts the data, and builds a clean archive.
The Hardware Requirements
To build this autonomous workforce, you need a specific setup.
It requires zero coding, but it does require the right tools.
→ The Claude Desktop app installed on your machine.
→ The Claude mobile app installed on your phone.
→ An active Pro or Max plan subscription.
→ Your computer must be powered on and awake.
This last point is critical.
If your laptop goes to sleep, your employee goes to sleep.
Keep it awake.
The 4-Minute Setup Guide
You don't need to know Python. You don't need terminal skills.
You just need to flip a few switches.
→ Download or update the Claude Desktop app.
→ Open the Cowork interface.
→ Click the "Dispatch" tab in the main menu.
→ Toggle "Local File Access" to ON.
→ Toggle "Computer Use" to ON.
→ Open your mobile app and start messaging.
Setup time is exactly 4 minutes.
Zero debugging required.
Dispatch vs. Regular Cowork
People confuse Cowork and Dispatch.
They are not the same thing.
Cowork is desktop-only.
It is session-based.
When you close the window, the session dies.
You have to be at your keyboard to use it.
Dispatch is fundamentally different.
It is cross-device. It is persistent. It is completely autonomous.
You can delegate a 45-minute contract review task and put your phone in your pocket.
You can walk away.
It triggers a massive mental shift.
From conversing to delegating.
Your desktop is no longer a passive tool waiting for your input.
Your desktop is now an employee that takes orders via text message.
Before You Continue: Why We Verify Subscribers
Quick note.
We have had multiple instances of bots scraping this content and people republishing it without authorization.
This guide took 40+ hours to build. It is free. But it is not public domain.
To keep the full playbook available for real humans (and out of the hands of content thieves), we require a valid email address to unlock the remaining sections.
It takes 3 seconds. No spam. No upsells. Just proof you are a real person.
Enter your email below to unlock the full guide.
Part 3: Revenue & Sales Frameworks
Sales is a numbers game. But tracking those numbers eats up your day.
Stop clicking through dashboards. Let your AI Chief of Staff bring the data to you.
Framework 1: The Revenue Pulse
When to use: Daily or weekly revenue check.
Open our payment dashboard. Pull the MRR, churn rate, and new customer count for the last 7 days. Compare these numbers to the previous 7-day period. Flag any anomalies or sudden drops. Deliver a one-page executive MRR snapshot.
What Claude Does:
→ Accesses your payment dashboard via computer use
→ Pulls the raw financial numbers
→ Compares current metrics to past periods
→ Builds a concise executive summary
What You Get Back: A clean revenue snapshot with MRR, churn, new customers, and red flags.
Pro Tip: Schedule this as a daily 7 AM routine. Wake up knowing exactly where your cash sits.
Framework 2: The Pipeline Surgeon
When to use: Weekly pipeline review.
Open the CRM. Audit all active deals in the pipeline. Identify stalled opportunities that have had zero activity in the last 7 days. Draft a short, personalized re-engagement email for each stalled deal based on the last recorded communication.
What Claude Does:
→ Opens your CRM and scans deal stages
→ Identifies stalled accounts
→ Reads past deal notes
→ Writes personalized follow-ups
What You Get Back: A list of at-risk deals with ready-to-send follow-up emails.
Pro Tip: Pair with the Follow-Up Machine for automated re-engagement sequences.
Framework 3: The Proposal Builder
When to use: New client opportunity.
We have a new client opportunity with [Client Name]. Search my email and Slack for all prior communication with them. Pull the meeting notes. Draft a structured proposal and Statement of Work based on what we discussed.
What Claude Does:
→ Searches your email and Slack history
→ Pulls relevant context and meeting notes
→ Drafts a structured proposal document
What You Get Back: A formatted proposal document ready for your final review.
Pro Tip: Save your best proposals as templates for Claude's memory. It learns your style fast.
Framework 4: The Pricing Auditor
When to use: Quarterly pricing review or before a major pitch.
Research the pricing pages for [Competitor 1], [Competitor 2], and [Competitor 3]. Compare their features and tiers to ours. Build a comparison matrix spreadsheet. Recommend specific adjustments to our pricing based on gaps.
What Claude Does:
→ Uses the browser to visit competitor sites
→ Extracts pricing data and feature lists
→ Builds a comparison matrix
→ Identifies gaps in your current model
What You Get Back: A pricing comparison spreadsheet plus strategic recommendations.
Pro Tip: Run this before every enterprise pitch. You will know exactly how to position your cost.
Framework 5: The Churn Predictor
When to use: Monthly churn prevention.
Scan our recent support tickets, product usage patterns, and payment history. Flag any accounts showing signals of churning. Cross-reference these data points and provide a ranked list of at-risk accounts with recommended interventions.
What Claude Does:
→ Cross-references multiple data sources
→ Scores account risk based on behavior
→ Identifies specific warning signs like dropped usage
What You Get Back: A ranked list of at-risk accounts with specific reasons and recommended actions.
Pro Tip: Set this as a weekly scheduled task. Catch the bleeding before it starts.
Part 4: Content & Marketing Frameworks
Attention is currency. But creating content from scratch every day leads to burnout.
Your AI Chief of Staff acts as your personal media agency. Treat it like one.
This is how I manage content across LeadPanther and my personal LinkedIn (52K+ followers) without a content team.
Framework 6: The Content Autopilot
When to use: Weekly content planning.
Analyze my top 10 performing posts from the last 30 days. Identify patterns in the topics and formats. Generate next week's content calendar. Draft 5-7 post concepts with strong hooks based on what already worked.
What Claude Does:
→ Pulls your post analytics
→ Identifies winning formats and topics
→ Drafts 5-7 post concepts with proven hooks
What You Get Back: A complete weekly content calendar with topics, hooks, and posting times.
Pro Tip: Schedule this every Friday at 3 PM. Walk into the weekend with next week's marketing done.
Framework 7: The Repurposing Engine
When to use: After a post gets strong engagement.
This post performed incredibly well: [Link/Text]. Repurpose this core idea into four new formats: a Twitter/X thread, a newsletter section, a LinkedIn carousel outline, and a 60-second short-form video script.
What Claude Does:
→ Analyzes the winning post structure
→ Adapts the tone for each specific platform
→ Formats the copy for different audiences
What You Get Back: 4-5 platform-specific content pieces from one original post.
Pro Tip: Run this within 24 hours of a post going viral. Momentum matters. Keep feeding the algorithm.
Framework 8: The Engagement Radar
When to use: Daily morning routine.
Scan the comments on my last 5 posts. Identify high-value commenters like founders or decision-makers. Draft a short, personalized reply to their comment. Also, draft a brief DM opener to continue the conversation in private.
What Claude Does:
→ Opens LinkedIn and reads comment threads
→ Researches commenter profiles
→ Writes contextual replies and DM openers
What You Get Back: A list of high-value commenters with drafted replies and DM scripts.
Pro Tip: Focus entirely on commenters who asked questions. They are your warmest leads.
Framework 9: The Competitor Content Tracker
When to use: Weekly competitive intel.
Go to the LinkedIn profiles of [Competitor 1], [Competitor 2], and [Competitor 3]. Pull their latest posts. Analyze their topics, hooks, and engagement levels. Identify content gaps that we can exploit in our own marketing.
What Claude Does:
→ Visits competitor profiles
→ Catalogs recent posts and scores engagement
→ Maps out their content strategy
What You Get Back: A competitor content audit with specific gaps you can exploit today.
Pro Tip: Track the same 5 competitors for 4 weeks straight. You will start to see their playbook.
Framework 10: The Lead Magnet Drafter
When to use: Creating a new lead magnet.
I am creating a lead magnet about [Topic] for [Target Audience]. Read these 3 source articles: [Links]. Produce a detailed outline. Once approved, write a full first draft section by section. Format it for distribution.
What Claude Does:
→ Reads the provided source material
→ Builds a logical structure
→ Writes the draft section by section
→ Formats it for easy reading
What You Get Back: A complete first draft of a 3,000+ word guide.
Pro Tip: Always provide source material. AI writing without sources is just AI guessing.
Part 5: Inbox & Communication Frameworks
Your inbox is a to-do list created by other people.
Stop letting it dictate your day. Let your AI Chief of Staff act as the gatekeeper.
Framework 11: The Inbox Zero Protocol
When to use: Daily morning routine.
Scan my inbox. Categorize all unread emails into Urgent, Respond Later, FYI, or Archive. Draft short, direct replies to the urgent ones based on my previous communication style. Archive the promotional noise.
What Claude Does:
→ Opens your email client
→ Reads subjects and body copy
→ Triages by priority
→ Drafts contextual replies
What You Get Back: A triaged inbox with drafted replies ready for your final approval.
Pro Tip: Schedule this at 6:30 AM. Your inbox is perfectly clean before you even pour your coffee.
Framework 12: The Slack Firefighter
When to use: After being offline for 2+ hours.
Scan all my Slack channels. Surface any threads where I am explicitly mentioned or where someone is waiting on my approval. Draft brief responses to unblock my team. Ignore casual chatter.
What Claude Does:
→ Reads channel history across your workspace
→ Identifies threads needing your input
→ Writes contextual, unblocking responses
What You Get Back: A summary of hot threads plus pre-written responses.
Pro Tip: Run this after every block of deep work or back-to-back meetings.
Framework 13: The Cold Email Responder
When to use: Inbound lead emails.
Review this inbound email from [Name]. Research the sender on LinkedIn and their company website. Assess if they match our Ideal Customer Profile. Draft a personalized response referencing their specific situation.
What Claude Does:
→ Searches for the sender's background
→ Evaluates if they match your ICP
→ Writes a reply that speaks to their specific pain points
What You Get Back: A personalized reply draft plus a quick profile summary of the sender.
Pro Tip: Speed kills deals. Run this within 1 hour of receiving the inbound email.
Framework 14: The Meeting Prep Brief
When to use: 30 minutes before any call.
I am meeting with [Name] from [Company]. Provide a one-page briefing with their background, recent company news, shared connections, and three talking points based on our past email history.
What Claude Does:
→ Researches across LinkedIn, company sites, and news
→ Scans past emails and meeting notes
→ Compiles a scannable brief
What You Get Back: A one-page meeting prep document.
Pro Tip: Include the phrase "Find anything they've said publicly about [Your Industry]" for instant rapport.
Framework 15: The Follow-Up Machine
When to use: After meetings.
Check my calendar for all meetings in the last 48 hours. Cross-reference with my sent email folder to see who I haven't followed up with. Draft follow-up emails for any gaps, including proposed action items based on the meeting titles.
What Claude Does:
→ Scans your calendar events
→ Checks your sent folder
→ Identifies the missing follow-ups
→ Writes personalized emails with action items
What You Get Back: Draft follow-up emails for every meeting that slipped through the cracks.
Pro Tip: Schedule this as a daily 5 PM routine. You will never drop a follow-up again.
Part 6: Client Management Frameworks
Your clients pay for results, not your time.
Manage them like a machine. Stop letting emails slip through the cracks.
I manage 5+ DFY lead magnet clients using these exact frameworks. LeadPanther captures the leads. These frameworks manage the relationships.
Framework 16: The Client Health Scorecard
When to use: Weekly client check (especially with 3+ active clients)
Scan all client Slack messages, emails, and meeting notes from the last 7 days. Score each client green, yellow, or red based on communication sentiment, response time, and open action items.
What Claude Does:
→ Reads across communication channels per client.
→ Analyzes sentiment and responsiveness.
→ Builds a traffic-light dashboard.
What You Get Back: A one-page client health dashboard with red/yellow/green scores and specific reasons.
Pro Tip: Red = call them today. Yellow = check in this week.
Framework 17: The Deliverable Tracker
When to use: Mid-week project status check
Scan project folders and check what deliverables are pending vs completed. Identify any items past deadline.
What Claude Does:
→ Opens local project folders.
→ Cross-references with task lists and deadlines.
→ Identifies missing assets.
What You Get Back: A status report showing exactly what's done, pending, and late.
Pro Tip: Add a constraint like "only flag items more than 2 days overdue" to filter out the noise.
Framework 18: The Onboarding Packet
When to use: The minute a new client signs
Client name is [Name], industry is [Industry], service type is [Service]. Compile a welcome doc with intake questions, a project timeline template, and a communication preferences form.
What Claude Does:
→ Pulls from your templates folder.
→ Customizes for the client's industry.
→ Builds a branded packet.
What You Get Back: A complete onboarding packet ready to send.
Pro Tip: Build once, iterate forever. Claude remembers your best templates.
Framework 19: The QBR Builder
When to use: Quarterly business review prep
Pull 90 days of communication, deliverables, metrics, and wins for [Client Name]. Draft a structured quarterly business review presentation.
What Claude Does:
→ Aggregates 3 months of data across email, Slack, and project files.
→ Builds a structured QBR deck.
What You Get Back: A presentation with metrics, wins, challenges, and next-quarter recommendations.
Pro Tip: Schedule this on the 25th of the last month of each quarter.
Framework 20: The Escalation Detector
When to use: Scheduled daily or every 48 hours
Scan all client communications from the last 48 hours. Flag any messages with frustration signals, repeated requests, negative language, or escalation phrases like "still waiting."
What Claude Does:
→ Runs NLP analysis across client channels.
→ Identifies tone shifts and repeated complaints.
What You Get Back: An early warning alert with flagged messages and recommended responses.
Pro Tip: This single framework prevents 90% of surprise cancellations.
Part 7: Operations & File Management Frameworks
Admin work kills momentum.
Let your AI handle the digital paperwork.
Framework 21: The File Organizer
When to use: Weekly desktop cleanup
Scan my Downloads and Desktop folders. Sort files into project folders by type (PDFs, images, spreadsheets, docs). Rename files with consistent naming conventions.
What Claude Does:
→ Reads file names and metadata.
→ Creates project subfolders.
→ Moves and renames files.
What You Get Back: A clean, organized file system with a summary of what moved where.
Pro Tip: Set a naming convention rule (e.g., "YYYY-MM-DD_ClientName_Description") in Claude's memory.
Framework 22: The Data Consolidator
When to use: Monthly reporting or any multi-source analysis
Pull data from [Source 1], [Source 2], and [Source 3]. Merge into one clean spreadsheet with consistent formatting.
What Claude Does:
→ Opens multiple files and screenshots.
→ Extracts relevant data.
→ Normalizes formats into a unified spreadsheet.
What You Get Back: One clean, merged spreadsheet with source attribution.
Pro Tip: Specify column headers in your delegation text so Claude matches your format exactly.
Framework 23: The SOP Writer
When to use: Right after completing a new process
Here are the rough steps for the process I just completed: [Dump steps]. Formalize this into a standard operating procedure. Include tool references, screenshot locations, and decision trees.
What Claude Does:
→ Structures your brain dump.
→ Creates numbered instructions.
→ Accounts for edge cases.
What You Get Back: A professional SOP document ready to hand to a team member or VA.
Pro Tip: "Write this for someone who has never done this task before" — forces Claude to be thorough.
Framework 24: The Invoice Processor
When to use: Monthly bookkeeping
Scan my inbox for invoices received this month. Extract the vendor name, amount, date, and category. Compile this data into a tracking spreadsheet.
What Claude Does:
→ Searches email for invoice messages.
→ Opens attachments to extract data fields.
→ Formats the numbers.
What You Get Back: A monthly invoice tracking spreadsheet with all payables organized.
Pro Tip: Schedule this on the 1st of each month as an automated routine.
Part 8: Strategic & Weekly Review Frameworks
You need to work on your business. Not just in it.
Use these to maintain altitude.
Framework 25: The Weekly Review
When to use: Sunday evening or Monday morning
Compile this week's metrics across revenue, content performance, and client status. List wins and blockers. Generate my top 3 priorities for next week.
What Claude Does:
→ Aggregates data from multiple sources.
→ Identifies patterns and bottlenecks.
→ Drafts a structured weekly review.
What You Get Back: A one-page weekly review document.
Pro Tip: This is your most valuable scheduled task. Set it for Sunday at 7 PM.
Framework 26: The Decision Memo
When to use: Any time you face a strategic decision
Here is the decision context: [Context]. Research both sides. Present pros and cons, make a recommendation, and include a risk assessment.
What Claude Does:
→ Researches the decision space.
→ Structures logical arguments.
→ Quantifies tradeoffs.
What You Get Back: A structured decision memo with a recommendation and risk analysis.
Pro Tip: Include "assume I have [budget/timeline]" to force practical recommendations.
Framework 27: The Competitive Intel Brief
When to use: Weekly or before a strategic planning session
Scan the websites, LinkedIn profiles, and public announcements from the last 7 days for these 5 competitors: [List]. Summarize their moves and threats.
What Claude Does:
→ Uses the browser to research competitors.
→ Identifies product launches, pricing changes, and hiring signals.
What You Get Back: A competitive intel briefing with specific moves and strategic implications.
Pro Tip: Track the exact same competitors for 4 weeks to spot hidden patterns.
Framework 28: The Goal Tracker
When to use: Monthly progress check
Here are my quarterly OKRs/goals: [List]. Check progress against each one, calculate completion percentage, and flag what's off-track with specific reasons.
What Claude Does:
→ Reviews goals against actual results.
→ Calculates progress percentages.
→ Identifies blockers.
What You Get Back: A goal tracking dashboard with completion percentages and course corrections.
Pro Tip: Pair with The Decision Memo when a goal is significantly off-track.
Part 9: Building Your Own Frameworks
You know the basics. Now you build your own system.
The Custom Framework Template
Don't start from scratch. Use this exact template to build your own delegation frameworks.
Role: [Who Claude needs to act as]
Context: [What is happening right now]
Task: [The exact output you need]
Constraints: [What NOT to do / formatting rules]
Sources: [Where to pull the data from]
Train The Memory
AI is only as good as its context. Train Dispatch on your business patterns.
Feed it your brand voice, your pricing sheets, and your standard operating procedures.
Tell it how you like your data formatted. It remembers. You never have to explain yourself twice.
The Rule of Three Delegations
If you do a manual task once, fine.
If you do it twice, take notes.
If you do it three times, build a framework.
Never repeat rote work. Turn it into a prompt. Let the machine handle it next time.
Automated Chains
Single prompts are good. Chains are better.
Combine your frameworks into automated sequences.
Run the Revenue Pulse, feed that into the Churn Predictor, and push the results through the Escalation Detector.
One click on Monday morning. Total visibility.
What Comes Next
You now have 28 frameworks that turn Claude Dispatch into a $150K Chief of Staff.
If you install 5 of these this week, you will reclaim 10+ hours.
If you install all 28, you will run your business like someone with an executive team.
The hard part isn't knowing WHAT to delegate.
It is building the discipline to delegate CONSISTENTLY.
Most founders read a guide like this, implement 2 frameworks, and then go back to doing everything manually.
Don't be that founder.
Here is how to keep the momentum going:
→ Join the Agent J community — 2,000+ operators building AI systems together. Weekly "Builders Hour" calls where we share frameworks, troubleshoot setups, and hold each other accountable.
Join Agent J on Skool
→ Check out LeadPanther — If LinkedIn lead generation is part of your business, LeadPanther automates the entire comment-to-DM-to-lead-magnet pipeline. It captures every lead from every post, automatically.
Visit LeadPanther.ai
Stop being the smartest person in the room doing the dumbest work.
Install these frameworks.
Delegate like a CEO.
Build like an operator.