How to Use AI in Your Business

This guide reflects how I use AI every day across marketing, sales and operations work, and what I recommend to clients. Tools change quickly, so treat the specific recommendations as a snapshot from late 2026 and the method as the part that lasts.

01 · The core idea

"Which AI is best?" is the wrong question

Most teams pick one AI and use it for everything. Results are inconsistent, and the conclusion becomes "AI isn't ready." The tool usually isn't the problem. The assignment is.

Choose the tool

ChatGPT, Claude, Gemini, Microsoft Copilot, Perplexity and Google AI Studio each have clear strengths. A research assistant is not a coding agent, and a coding agent is not a calendar assistant.

Choose the model and effort

Most tools let you select a model and how much reasoning it applies. The most capable setting belongs on complex, high-stakes work. Lighter settings handle routine tasks faster and at lower cost.

Think of it as staffing. You wouldn't assign your most senior specialist to scheduling, or ask a coordinator to make a strategic call. Match the work to the right resource.

02 · Effort levels

Match the effort to the stakes

Most AI tools now let you control how deeply the model reasons. A simple rule covers most decisions:

High effort Costly mistakes

Strategy, financial analysis, architecture decisions, complex code, and anything where an error is expensive to fix.

Standard effort Everyday work

Drafting, summarizing, organizing, formatting and routine edits. Fast, inexpensive and good enough.

The cheaper option isn't always cheaper. In my experience, a top model can complete a complex build correctly in minutes, while a lighter model can take hours of supervision and rework on the same task. Measure the total time to a correct result, not the price per request.

03 · Workflow patterns

Use more than one AI on the same job

The goal isn't to find the one best tool. It's to sequence tools so each handles the part it does best.

Pattern 1 · Build, then refine
A top model such as Claude Opus produces the structure: the plan, the core logic, the difficult part. A faster tool handles revisions, formatting and small follow-up changes. You get premium quality while reserving premium usage for where it matters.
Pattern 2 · Research, then decide
A research tool such as Perplexity or ChatGPT's browsing mode gathers current sources. A reasoning model then compares the options against your constraints and recommends one.
Pattern 3 · Standardize across tools
When you settle on a rule, a format or a process with one AI, have it write a short instruction note. Store it where every AI your team uses can read it, so results stay consistent across tools and people.

04 · Reference table

Which AI for which job

A starting point based on my own use as of late 2026. Re-test regularly; rankings shift every few months.

Job	Start with	Why
Strategy, planning, complex analysis	Claude or ChatGPT (top model, high effort)	Strongest multi-step reasoning
Coding and technical builds	Claude Code	Most precise coding agent I've used
Daily writing, summaries, email drafts	ChatGPT	Versatile, fast, generous limits
Research with current sources	Perplexity or ChatGPT search	Cites sources you can verify
Very large documents	Google AI Studio (Gemini)	Very large context window
Work inside Microsoft 365	Microsoft Copilot	Built into Outlook, Word, Excel and Teams
Work inside Google Workspace	Gemini	Built into Gmail, Docs and Sheets
Images and visual drafts	ChatGPT	Strong image generation and editing
Scheduled reports and recurring tasks	ChatGPT or Claude with scheduled tasks	Run on a timetable without prompting
Talking through data hands-free	Gemini voice	Fluid voice conversation

Use the most capable setting for high-stakes work and a standard setting for everything else. That one habit fixes most AI disappointment.

05 · Prompt library

Tested prompts for daily operations

These are prompts I use repeatedly, rewritten as templates. Copy one, fill in the placeholders and send.

Planning and decisions

Turn a goal into a planLarge goals. Use high effort.

Turn the goal below into a practical execution plan. First identify the outcome, constraints, assumptions and missing information. Only ask questions that would change the plan. Break the work into prioritized steps, each with a deliverable, dependencies and a definition of done. Flag the biggest risks and how to reduce them. End with the next three actions I can take this week.
Goal and context: [goal]

Get an honest assessmentBefore committing to an idea.

Here is an idea I'm considering: [idea]. Give me your honest assessment. What's strong, what's weak, what would make it fail, and what would you do differently? Don't default to the conventional answer if the evidence points elsewhere.

Recommend one optionWhen you need a decision, not a list.

What is the best [tool / vendor / approach] for [situation and constraints]? Give me one recommendation and explain why. Then name the runner-up and when I'd choose it instead.

If you were in chargeStrategic options. Use high effort.

If you were responsible for [goal] with [budget, team, timeline], how would you approach it? Give me two or three ranked approaches with the trade-offs of each.

Meetings and communication

Clean up a transcriptAfter calls or voice notes.

Clean up this transcript. Fix transcription errors and add paragraph breaks. Don't shorten, summarize or change my wording.
{transcript}

Extract action itemsAfter any meeting.

From this transcript, list every action item, decision and open question. For each action item, include the owner and due date if mentioned. Keep the original wording where possible.
{transcript}

Debrief a sales callAfter client conversations.

Here are my notes from a call with {person / company}: {notes}. Give me: 1) the three things that matter most, 2) what they need versus what they said, 3) objections and how to address them, 4) a follow-up email draft (don't send), 5) the most valuable next step.

Write a prompt for another AIWhen handing work to another tool or person.

Turn this request into a clear brief I can give to another AI or a team member: [request and context]. Include the goal, sources, constraints, deliverables and how to verify the work is done.

Create a handoffWhen switching tools or people mid-project.

Write a handoff for [project]. Include the current situation, goals, decisions made, where files live, constraints, work completed, open issues and next actions, so someone can continue without redoing the research.

Research

Compare options for my situationVendor, tool or product decisions.

Find [product / service / tool] options that fit [budget, size, requirements]. Research current offerings, verify pricing and key details from official pages, compare the top three, and include direct links. Explain why each does or doesn't fit my situation.

Verify before answeringWhen accuracy matters.

Answer this using my own records first: [files, emails, reports]. Only use outside sources to fill gaps, and label which is which. If you can't verify something, say so instead of guessing.

Data and systems

Read or update a database safelyAirtable, spreadsheets, CRMs.

Here is my [Airtable base / spreadsheet / CRM]: [link]. I want to [log / update / organize] this: [details]. Inspect the existing fields and records first. Don't create duplicates or overwrite existing data. Show me the changes before saving them.

Review a processBefore changing how something works.

Here is how we currently handle [process]: [description]. Identify bottlenecks, unnecessary steps and what could be automated. Recommend the three changes with the highest impact for the least effort. Don't change anything yet.

Reconcile spendingMonthly finance review.

Compare these charges against our budget and recurring expenses: {statement}. Flag anything unexpected, duplicate or no longer needed, list payments due in the next 14 days, and summarize total spend by category.

Client messages

These aren't AI prompts. They're replies I use with clients and prospects.

Checking inAfter a conversation.

Hi {name}, following up on our conversation about {topic}. Are you still looking to move forward? If the timing has changed, that's completely fine. Just let me know where things stand.

No responseAfter a few weeks of silence.

Hi {name}, I haven't heard back, so I'm guessing priorities have shifted. Should I close this out on my end, or is it still on your list?

Pricing questionBefore you understand the scope.

Good question. It depends on what we're solving, and I'd rather give you an accurate number than a guess. Can we schedule 15 minutes so I can understand the scope and give you a real figure?

Price objectionReframe against the outcome.

That's fair. Compared to the cost of the problem, though: if this delivers {result}, how does the investment look? If it still doesn't fit, we can start with the piece that pays for itself first.

Needs time to decideFind the real concern.

Of course. Which part are you weighing most: the investment, the timing or confidence that it will work for you? I'm happy to answer any of those directly.

Next stepClose the loop.

Thank you, {name}. The next step on my side is {next step}, and you'll have it by {day}.

By Ron Pascal

Ron Pascal has 12+ years in marketing and sales and has worked with more than 500 clients. He publishes the playbooks he uses with clients.

Contact Ron More about me →

Next guideChatGPT and Claude Usage Limits: How to Get More From Every PlanWhy teams hit AI usage limits, and the routing, folder and context habits that make every plan go further.

Newsletter

Get new guides by email.

New guides and articles when they’re published. No spam, and you can unsubscribe anytime.