Dharmesh Shah’s Expert Tips for Working with Claude and ChatGPT
HubSpot’s co-founder and CTO shares his battle-tested strategies for maximizing AI productivity
When the co-founder of a $20+ billion company spends his nights coding AI tools and building agent platforms, it’s worth paying attention to his advice. Dharmesh Shah, HubSpot’s CTO and the mind behind ChatSpot and Agent.ai, has been experimenting with AI since before ChatGPT existed. Here are his proven strategies for getting the most out of Claude and ChatGPT.
The Mindset Shift That Changes Everything
Shah’s research revealed a fundamental divide in how people approach AI: “About a third of the people thought about that kind of negative scenario, I’m competing against AI, and two thirds thought they were competing using AI.” Dharmesh Shah on LinkedIn: Not sure if this is new, or I just missed it, but ChatGPT now shows… | 46 comments The winners? Those who see AI as a collaborator, not a competitor.
Shah’s Golden Rule: “Ask yourself, any time you do something, you’re sitting at the computer about to start something, ask yourself, ‘How might I use AI to help me with this?’ Put aside the skepticism. Just type something into ChatGPT or your favorite tool of choice and see what happens.” Dharmesh Shah on LinkedIn: Not sure if this is new, or I just missed it, but ChatGPT now shows… | 46 comments
Programming in the Language of the Future: English
One of Shah’s most powerful insights is reconceptualizing how we interact with AI. He describes someone using a 2000-word prompt to build fantasy fiction worlds: “Basically, what he’s doing is he’s programming in what will be the most popular programming language in the world someday, English, right? He’s just giving the computer instructions as to what he wants.” Dharmesh Shah on LinkedIn: Not sure if this is new, or I just missed it, but ChatGPT now shows… | 46 comments
This isn’t just clever wordplay—it’s a fundamental shift in how we should approach AI prompting. Think of yourself as a programmer, but instead of Python or JavaScript, you’re coding in natural language.
The Natural Language Revolution
Shah explains why this matters: “We’re taking the mental model that’s in our head of what we’re trying to accomplish with said piece of software and translating that into a series of touches and swipes and clicks and things like that. And there’s nothing natural or intuitive about it.” The Agent Network — Dharmesh Shah – Latent.Space
With AI, you can finally skip the translation layer. Express your thoughts directly, and let the AI figure out the execution.
Real-World Examples That Work
Shah’s ChatSpot tool demonstrates the power of clear, direct prompting. Here are examples from his actual implementation:
- Data Analysis: “How many contacts in the CRM are from California?”
- Lead Generation: “Find companies in the food industry in Florida with over 250 employees”
- Content Creation: “Draft a professional email to a prospect in the automotive industry”
- Visualization: “Show me my monthly website traffic for last year as a bar chart”
- Reporting: “Create a report of companies added in Q4 summarized by country”
Notice the pattern? These prompts are:
- Specific and actionable
- Use business language, not technical jargon
- Request concrete outputs
- Combine multiple data sources seamlessly
Advanced Integration: The MCP Game-Changer
Shah is particularly excited about Model Context Protocol (MCP), calling it a “universal connector” for AI and data. With Claude Desktop and MCP: “I can have the LLM use agents on Agent.ai, access CRM data in HubSpot, read/write to a specified directory in my local file system, read/write messages to Slack, and access my Google Calendar and Gmail.” HubSpot’s Founder Predicts the Next Big Thing In Enterprise AI – CX Today
This represents the future of AI interaction—seamless integration across your entire digital workspace.
The Art of Follow-Up Questions
Shah appreciates AI features that suggest next steps, noting that when exploring new subjects, “you don’t know what you don’t know.” His recommendation? Don’t stop at the first answer. Ask:
- “Is there anything else I should know about this?”
- “What other questions might be relevant to this subject?”
- “What are the implications of this for [your specific context]?”
Learning Through Experimentation
Shah’s approach to mastering AI is refreshingly simple: “Just start learning by doing it and give it a shot and you’ll be surprised how often it works.” Dharmesh Shah on LinkedIn: Not sure if this is new, or I just missed it, but ChatGPT now shows… | 46 comments
The key is overcoming analysis paralysis. Start experimenting today, not tomorrow.
Beyond Simple Prompting: Creating Real Value
Shah warns against building “thin wrappers around the GPT APIs.” Instead, focus on:
- Solving real problems
- Combining AI with unique data or insights
- Creating compound value through integration
- Building workflows, not just individual queries
The Human-AI Partnership Philosophy
Perhaps Shah’s most profound insight: “The better our artificial tools get, the better AI gets, the more human it allows all of us to be. We’re spending more time on the things that make us human because human is not a bug. It is the ultimate feature that we all have.” Dharmesh Shah on LinkedIn: Not sure if this is new, or I just missed it, but ChatGPT now shows… | 46 comments
Action Steps for Your Business
Based on Shah’s approach, here’s how to implement these strategies:
- Start Small: Pick one repetitive task this week and ask, “How might AI help with this?”
- Think in Workflows: Don’t just automate tasks—redesign entire processes around natural language interaction
- Experiment Daily: Spend 15 minutes each day testing new prompts and approaches
- Integrate Gradually: Look for ways to connect AI tools with your existing systems
- Focus on Value: Always ask whether you’re creating genuine value or just impressive demos
The Bottom Line
Dharmesh Shah’s success with AI isn’t because he’s a technical genius (though he is). It’s because he approaches AI with curiosity, persistence, and a focus on real-world value creation. His message is clear: the future belongs to those who learn to collaborate with AI, not compete against it.
Start where you are. Use what you have. Do what you can. The AI revolution isn’t coming—it’s here, and the best time to start experimenting was yesterday. The second-best time is now.
Ready to transform your business with AI? Start by implementing one of Shah’s strategies this week and see the difference it makes. The tools are here—the question is, will you use them?