Founder and CTO at HubSpot. Helping millions grow better.
Boston, Massachusetts, United States
1M followers
500+ connections
1M followers
500+ connections
About
Career startup guy.
Author of the book "Inbound Marketing" (http://InboundBook.com)
Specialties: Entrepreneurship, Software Development, internet marketing,SEO
Articles by Dharmesh
Activity
1M followers
-
Dharmesh Shah posted thisAI is getting better and better at answering CAPTCHAs. Me? I'm failing them more often. Soon, I'm going to have to start using AI to prove that I'm human.
-
Dharmesh Shah posted thisFDEs (or Forward Deployed Engineers) are all the rage these days within tech companies. The idea is simple: To help customers start getting real value really quickly, send out an engineer that can help with implementation, customization, optimization, etc. (I'm using "send out" figuratively -- they may not need to go on-site). Not sure when this started, but Palantir popularized it recently. But FDEs don't have to be limited to just engineers. I like to think of FDEs as "Forward Deployed Experts". What you really want is people that are deep on the thing that needs to be done for customers to get value. Sometimes its an engineer. But it could be an AI-forward lawyer/broker/teacher/consultant/whatever (based on the industry you're in). I think this has the potential to open up a bunch of new career opportunities for those that are curious, competent and obsessed with solving customer problems.
-
Dharmesh Shah shared this"The first rule of Dunning-Kruger Club is you don't know you're a member of Dunning-Kruger Club." ~ David Dunning I'm fascinated by cognitive biases and the Dunning-Kruger effect is one of them. The Dunning-Kruger Effect is a cognitive bias where individuals with limited knowledge or progress in a domain significantly overestimate their abilities, while experts tend to underestimate theirs, often unknowingly. I think we have a similar effect going on with AI right now. AI tools amplify our abilities. They are awesome (I'm a big fan and daily beneficiary). But, it's easy to fall into the trap that because we have access to this power, that we all of a sudden become expert geniuses in a given discipline or domain. I'll call it the Dharmesh AI Delusion Disorder. Just because I use my image generation tool (quick plug for ImageGen .ai) to create really awesome designs, all in a consistent style does not turn me into a world-class designer. Being able to make the artifact is not the same as having mastered the art. p.s. My hope is that once one figures out one is in the club, one is no longer in the club. :)
-
Dharmesh Shah posted thisCurrent Status: Working on a major update to HubCode (The agentic coding tool for building HubSpot apps). The issue I (and other users) were running into was the 15 second timeout limit when using the fetch() function to call external endpoints. When originally implemented, the 15 second timeout made sense because this was before the age of AI, and calls were deterministic and generally just looking up some external data. 15 seconds was plenty (typical calls took a few seconds). But now, in the age of AI, we often want our apps calling endpoints that are longer running -- because they're using one more LLM calls, invoking agents and doing other things that take longer than 15 seconds. So, I am thrilled that the HubSpot team, within a matter of days, rolled out an update so that you can (optionally) increase the timeout period to 120 seconds. Docs have been updated. This opens up a bunch of use cases around invoking Agent.ai agents from within app cards and apps. Testing things out as we speak. If you haven't signed up to try it, visit HubCode .com (there's a beta list, but I'm cranking through it and approving pretty quickly).
-
Dharmesh Shah posted thisWe used to live in a world where we'd sometimes be up at 3am, questioning our life choices. Now we we're up at 3am questioning our LLM choices.
-
Dharmesh Shah posted thisI'm going to say something shocking and borderline heretical (for me). I'm thinking about switching from Python to TypeScript for my next project. I've been using Python for *decades* as my primary programming language. There are so many things I love about the language: 1) Python is a joy to develop in. 2) Great libraries for everything I need to do. 3) It seems to be lingua franca of AI. It's what LLMs use when they need to generate code to solve a problem. It's usually the first language to get an SDK/library from the frontier model companies (and others in the space). It's what many AI-oriented open source projects are built-in. So, why am I considering TypeScript? A few reasons: 1) There's elegance and value in having both my front-end code and back-end code in the same language. 2) TypeScript is natively type-safe. 3) When distributing applications to others (particularly CLI tools), it's much more common/simple to do with a Node app then try to build binaries in Python (using something like pyinstall). 4) TypeScript is a close second when it comes to being popular in the AI community. 5) There are packages for most of the common things I need (web framework, database access, web/http calls, etc.). 6) There is first-class support from Vercel for deploying TypeScript apps. I'm both an investor in Vercel and a customer) but have mostly used it for front-end deploys, not backend. And, what once was the biggest reason NOT to use TypeScript is no longer true: The fact that I don't know TypeScript and didn't want to spend hundreds of hours becoming an expert at it. Now with agentic coding, I don't need deep knowledge of the language in order to be productive. With my knowledge of Python, C++ and other prior languages, I can likely get by pretty well in TypeScript with the help of Codex and Claude Code. Haven't made the decision to switch over completely yet, but next time I have a small, contained project I need to work on, I'm considering trying TypeScript. What do you think? Am I overthinking it or underthinking it?
-
Dharmesh Shah posted thisImagine this: LinkedIn lets you create a prompt/instructions (let's call it MESSAGES.md and INVITES.md). You can write instructions as to how to handle DMs and invites to connect. Examples: 1) Auto-archive any messages that are requests for a "paid expertise" project. 2) If it's a message or invite from a YC startup founder, prioritize it. 3) if I get an invite to connect from someone that I have zero overlapping connections with, auto-archive it. 4) If someone messages about an invite to speak at a conference, auto-respond with "Thank you for the invitation, but I'm not accepting any speaking engagements for the remainder of 2026. Best wishes with your event." You get the idea. Use AI for something that it is really good at (analyzing and classifying) so I don't have to do the thing I'm really bad at (following up on invites and DMs). This would be completely and totally game-changing for my use of LinkedIn. And, this feels like it would be relatively straight-forward to implement and would not compromise member data in any way (since there's no third-party application involved). What do you think? Would you use this if it were available? Would you pay for it?
-
Dharmesh Shah posted thisClosed-loop > closed-won. We all love closed-won deals. We signed up a customer. Yay! But what's even better than getting a closed-won deal? Using that context to close the loop so the underlying system gets smarter and sparks future deals. AI models are genius-level smart. But there's a limit to the the value they can generate based on their genius alone. To really hit breakthroughs and drive *future* growth, we need to give AI the context made possible by closed-loops. It's why I'm so excited about HubSpot's Agentic Customer Platform. It doesn't just capture the data, it synthesizes the context for use in the future by AI. cognition + context = compounding growth
-
Dharmesh Shah shared thisCURRENT STATUS: Working on a presentation to one of the largest audiences I will speak to this year. The global HubSpot team (~9,000 people). We are hosting what we call GrowDAI where we spend a day in a mini "conference" learning about AI. The content is by HubSpotters for HubSpotters. Yes, we love AI, but we love people even more and want them to learn and grow. Looking for topic ideas or questions that you think might be of interest (the hardest part of these presentations is figuring out what to say). Even if you're not on the HubSpot team (yet) imagine that you are. What would you want to learn about? What are you curious about? If you were moderating a fireside chat, what would you ask? I'll do my best to post an edited version of the video with the publicly shareable parts. Thanks for your support.
Dharmesh Shah commented on a post
13m
Happy Anniversary! Thanks for all that you do.
-
Dharmesh Shah reacted on thisDharmesh Shah reacted on this10 years at HubSpot this week! Still a little surprised by that. When I first showed up to join the customer support team, I was honestly just looking for a job to start my life after graduate school. I doubted I'd stay for more than a year or so, & I wasn't even sure how much I wanted to be a tech worker. It's pretty different from studying literature & gender politics. Never would've thought I'd still be here 10 years later, doing a job I love with a lot of wonderful people. I've been asked a few times recently what has kept me at the same place for so long, & more than anything it's the people: the connections I've made with others who I've learned from & been able to consistently grow alongside. I do a lot of work I'm very proud of, & I'm fortunate to be at a place with a mission I believe in & where customers come first, which makes loving the job easy. But all the folks I've worked with over the years have left a mark on me much larger than anything else. If we've crossed paths at HubSpot, especially if we've worked closely together, thanks for making the last 10 years something special.
-
Dharmesh Shah liked thisDharmesh Shah liked thisEarlier this week, I wrote about how HubSpot has transformed how we build and go to market. But neither of those would have been possible without transforming something more fundamental: how we operate as a company. That’s what I want to share with you today. When we started this journey, we made a deliberate choice: don't mandate AI. Inspire it. We gave everyone access to tools, created a culture of experimentation, and replaced annual planning with six-week sprints. Now, 94% of HubSpotters use AI weekly, and some teams have dramatically increased their impact. But we’re just getting started. Helping employees become 10x more efficient does not automatically make a company 10x more productive. So now, we're focused on something deeper: driving institutional AI productivity. That all starts by building a foundation of shared context for every HubSpotter, every team, and every agent. You can read the full story here, including the results we've seen and where we're headed next: https://lnkd.in/eUc4aYGH P.S. This wraps up a series I've been sharing this week. Over the last three years, every corner of HubSpot has changed in some way. My guess is it’s the same for your company. If it’s not, then it will be soon. I hope this series helps you, whatever stage of the journey you’re on. The companies that successfully transform with AI will have a compounding advantage for years to come. I want every HubSpot customer to be one of them.
-
Dharmesh Shah liked thisDharmesh Shah liked thisJust finished "Generative AI for Sales and Services Professionals" by Sam Mallikarjunan! Check it out: https://lnkd.in/dY7vcjNw #artificialintelligence, #generativeaiforsales, #salessupporttools.
-
Dharmesh Shah liked thisDharmesh Shah liked thisI’m officially no longer a marketer. And I’ve moved into a role that likely didn’t exist 6 months ago. At HubSpot, we’ve been integrating AI across our GTM. This has been one of the initiatives I’ve been proud to lead - the Flywheel AI group. A cross-functional pod with concerted bets. We’ve had some good success: → 345,000 net new accounts added to our TAM in a year → 82% of inbound chats handled with zero humans → 1,850% growth in qualified leads from ChatGPT and Perplexity → 10,000+ meetings booked per quarter from AI-personalised outreach → 13% lift in win rate on deals using AI guidance → 60% of internal support inquiries resolved without a human → 7-point lift in customer save rate We can then pass those learnings back to customers via our product and playbooks. As the models have improved, we’ve realized the swings to take our larger. Instead of integrating AI across the GTM, you need to rebuild your GTM around AI. To do that, we’ve created a talented group of people across areas like product, engineering, data science, operations, marketing technology, digital experience, and more. We call it the Agentic GTM & Systems group. The premise is to move fast, take bigger bets, find signals, and pass learnings to customers in the form of product and playbooks. You can read our journey so far from Yamini Rangan: https://lnkd.in/esJvkR4X What’s fascinating is how new roles will evolve over time as GTMs get rebuilt. I’ll be sure to share everything I learn :)
-
Dharmesh Shah liked thisDharmesh Shah liked thisHubSpot has launched a prospecting agent that costs $1 per lead. Not per seat or month. Per result. This is one of the most interesting GTM pricing shifts since Slack went freemium. Every SaaS company spent the last 15 years building around seat-based pricing because it was predictable and easy to model. But AI agents don’t sit in seats. They produce outcomes. And the moment a major vendor prices on outcomes, buyers start doing the math: Why am I paying for seats when I could be paying for results? The downstream effect is what matters. If your AI agent delivers a qualified lead for $1 and a human SDR delivers one for $150, that’s a headcount conversation. And most RevOps aren’t ready to have it yet. I’ve seen two major shifts in SaaS pricing models. Both times, the companies that moved first looked expensive for about six months. Then they looked inevitable. I don’t know if outcome-based pricing will become the norm. But I do think the companies still forecasting around seats are going to feel this shift before they see it coming. And we need to think about what it means for the people currently in those seats. My bet is that the best ones become more valuable, not less. That’s humans doing what agents can't - feel the nuance, build trust, think laterally. They become the differentiator. Curious about your read. Is this the beginning of something, or a niche model that works for prospecting and not much else?
Experience
-
-
-
Founder and CEO
Pyramid Digital Solutions
-
Education
Licenses & Certifications
Skills
Publications
-
Inbound Marketing
United States
Projects
-
simple.ai
The AI newsletter that makes the complicated world of AI more simple.
-
agent.ai
Agent.ai is the professional network for AI agents.
Discover, build and engage AI agents. Almost 2 million (human) users and 1,500+ agents.
Languages
-
English
Native or bilingual proficiency
-
Gujarati
Native or bilingual proficiency
Recommendations received
7 people have recommended Dharmesh
Join now to viewOther similar profiles
Explore more posts
Explore top content on LinkedIn
Find curated posts and insights for relevant topics all in one place.
View top contentOthers named Dharmesh Shah in United States
55 others named Dharmesh Shah in United States are on LinkedIn
See others named Dharmesh Shah