Sign in to view Jesal’s full profile
or
New to LinkedIn? Join now
By clicking Continue to join or sign in, you agree to LinkedIn’s User Agreement, Privacy Policy, and Cookie Policy.
Sign in to view Jesal’s full profile
or
New to LinkedIn? Join now
By clicking Continue to join or sign in, you agree to LinkedIn’s User Agreement, Privacy Policy, and Cookie Policy.
San Francisco Bay Area
Sign in to view Jesal’s full profile
Jesal can introduce you to 9 people at cora.ai
or
New to LinkedIn? Join now
By clicking Continue to join or sign in, you agree to LinkedIn’s User Agreement, Privacy Policy, and Cookie Policy.
5K followers
500+ connections
Sign in to view Jesal’s full profile
or
New to LinkedIn? Join now
By clicking Continue to join or sign in, you agree to LinkedIn’s User Agreement, Privacy Policy, and Cookie Policy.
View mutual connections with Jesal
Jesal can introduce you to 9 people at cora.ai
or
New to LinkedIn? Join now
By clicking Continue to join or sign in, you agree to LinkedIn’s User Agreement, Privacy Policy, and Cookie Policy.
View mutual connections with Jesal
or
New to LinkedIn? Join now
By clicking Continue to join or sign in, you agree to LinkedIn’s User Agreement, Privacy Policy, and Cookie Policy.
Sign in to view Jesal’s full profile
or
New to LinkedIn? Join now
By clicking Continue to join or sign in, you agree to LinkedIn’s User Agreement, Privacy Policy, and Cookie Policy.
About
Welcome back
By clicking Continue to join or sign in, you agree to LinkedIn’s User Agreement, Privacy Policy, and Cookie Policy.
New to LinkedIn? Join now
Activity
5K followers
-
Jesal Gadhia shared thisWe’re hitting some big milestones at Cora. Proud of this team and grateful to our customers for sharing their stories!Jesal Gadhia shared thisWe crossed $3B in revenue managed by Cora agents just months after launch... so we hopped on a plane to hear from customers on the outcomes Cora is driving. Some of the highlights we heard: → HeyGen reactivated hundreds of dormant customers, with direct impact to net dollar retention → BetterUp saw 5,000 hours and at least a million dollars saved from automating manual work → Flock re-engaged thousands of long-tail customers with personalized touch
-
Jesal Gadhia shared this"Ownership" might be the most overused word in engineering. It's also one of the least understood. Most people hear "can you own this?" and think it means "will you do the task." Take the ticket, write the code, close it out. That's just doing your job. I don't usually repost other people's writing. This is the exception. Thorsten Ball captured the real version better than anything I've managed to say about it. Owning a problem means taking it from "we have a problem" to "we never have to think about it again." The task is the small part. The rest is everything nobody assigned you: - Working out what the actual problem is instead of jumping to the first solution. - Thinking through edge cases and failure modes before they turn into support tickets. - Knowing how you'll prove the thing works. - Telling the people it affects. - Checking the logs a week later to confirm it's still doing what you thought. None of that lives on the ticket. All of it is the work. And AI has made this more true, not less. The part it's best at is the part that used to count as the job. Writing the code is close to free now. What's left is everything around it, and almost all of that is still human. The gap between a good engineer and a great one was never raw output. Now it's the whole game. How wide a circle of responsibility you draw around a problem without being asked. That's what people mean by a 10x engineer. Not the ones who move fast. The ones who own something so completely that you never think about it again.
-
Jesal Gadhia posted thisEveryone's debating what is AI-native vs not. I think there's a simpler test. Can you go back? Not "would you want to." Can you actually imagine doing the work the old way again? I use Granola for every meeting. First week, I just tried it. By week two, I'd stopped taking notes entirely. Going back to a notes doc now feels like going back to pen and paper. That's what AI-native actually feels like. Not a chatbot. Not a memo about AI transformation. Not an architecture diagram. A quiet, irreversible shift in what normal feels like. That's what we're building with cora.ai. An AI that keeps tabs on all customers across your entire book, tells you the vibe of the account before you jump on a call, builds your QBRs, tracks renewals, chases down deliverables so nothing slips through, and more. Once you have that, can you go back to managing it all manually? Pulling engagement data, juggling spreadsheets, writing the same check-in emails, deciding which accounts to prioritize this week? That's the bar. Does using it change what you can tolerate going back to?
-
Jesal Gadhia posted thisFor most of my career, it was just me and a shovel. You get good at the shovel. You feel the ground, you know when you're hitting rock, you adjust. Then I found a bulldozer. Same job, more power. I was still the operator, still making every move. Just faster. Then the crew showed up. All of a sudden I had a jackhammer operator, an excavator operator, a team of people who knew their tools better than I ever could. My job stopped being about operating anything. It became about knowing what to build, in what order, and making sure the right equipment was pointed at the right problem. That's the arc of AI right now. Most people are somewhere between the shovel and the bulldozer, getting comfortable with the power. The agentic shift is something different. You're not the operator anymore. You're the site manager. The agents are making decisions. Your job is direction, sequencing, and the calls only a human should make. The people who struggle are the ones who never got comfortable with planning. The ones who thrive are the ones with mental clarity and a habit of thinking before doing. My old friend and colleague Kevin Mehlbrech wrote a sharp piece on the risks that come with that transition. Link in the comments.
-
Jesal Gadhia posted thisCoinbase announced this week that they're reorganizing around one-person pods, a single person carrying what used to require an engineer, designer, and a PM. This is where the industry is heading. And it raises the bar for what an engineer actually needs to bring. The model only works if that one person has what agents can't generate: judgment. Knowing which problem is worth solving. Knowing which output to reject. Knowing when to stop going down a rabbit hole and make a call. This is how we think about building at cora.ai. AI handles the execution layer. The human carries the context, the taste, and the critical thinking. I was featured in LeadDev last week on exactly this. Critical thinking ranked #1 when engineering leaders were asked which skills will matter most over the next three years. If you're an engineer who brings strong editorial judgment and taste to AI, we're hiring. Link in comments.
-
Jesal Gadhia posted thisConnect an AI to every tool your company uses. Gong, Gmail, Salesforce, Slack. Ask it what's happening with a customer account. It'll give you an answer. It'll sound confident. And it'll miss the point. We learned this building cora.ai. The AI could see everything. Every call transcript, every email, every CRM update. But it couldn't tell the difference between a VP skipping a meeting because they're busy and a VP skipping a meeting because they're evaluating your competitor. Access wasn't the problem. Judgment was. So we started building the judgment layer. What actually qualifies as a risk signal vs. noise. When a complaint is about your product vs. about an SLA. Which contacts matter and which are just cc'd on everything. Months of learning from real customer data, encoded into rules the AI can use every time, not just once. That's the part nobody talks about. Connecting to tools is easy. Every demo does it. The hard part is teaching the agent what to pay attention to and what to ignore. That's not a model problem. It's a knowledge problem. An agent with access reads the transcript. An agent with a customer context layer reads the account.
-
Jesal Gadhia posted thisSometimes we get asked if we support sending surveys. My first instinct is to say no. Not in the traditional sense. No form builder. No question templates. No NPS widget. But then we ask the question we ask before building anything: what's the agentic version of this? Our agents already chase deliverables from customers. They process responses and convert them to structured data. The agent compiles that into a handoff package. That's a survey. We just never called it one. This keeps happening. When you're building an AI-native product from the ground up, you don't have legacy features to protect. There's no existing survey module with a product manager and a roadmap. So when a customer asks for something, the question isn't "how do we add this to the product?" It's "what would the agent do if it had to solve this with the platform we've already built." Half the time, the answer is already there. Just not labeled the way the customer expects. You don't build a survey tool with form builders and dashboards. You build an agent that collects responses and feeds them back into the workflow. Less to maintain. More flexible. Works for the next use case without a feature request sitting in a backlog that never gets shipped. Traditional software companies are retrofitting AI onto architectures that were built for a different era. AI-native companies ask what the first principles way of solving this problem with an agent is. That reframe has changed how we build more than any technical choice we've made.
-
Jesal Gadhia reposted thisJesal Gadhia reposted thisPeople often ask me what an "AI-native" engineer looks like. The best answer I can give is it looks like Pat Nadolny. Pat was our first full-time hire and founding engineer. Believe it or not he found us on Hacker News when we were quietly building in stealth... and he couldn't have been a more perfect fit for cora.ai. He had built an AI data analyst at another startup, and his very first message to us was asking sharp questions about balancing agent autonomy with human-in-the-loop. We could tell he was a true builder to his bones, with the Github contributions and side projects to show for it (shoutout to his "Cursor for video editing"). Pat's a unicorn. Ex-consultant so he knows how to build trust with customers. Data engineering background so he can tackle the messy system design challenges behind the world's richest customer context graph. And he lives on the edge of AI tools... knows when to think slow vs where coding agents give him 10x leverage. But what stands out most is his instinct for the customer. When he's not building alongside them, he's spending his free time listening to sales call recordings for what's resonating and forming a point of view on where we need to go next. If you want to work with 10x engineers like Pat, we're hiring.
-
Jesal Gadhia posted thisEveryone's trying to build AI agents for the enterprise right now. Most of them are going to fail. And not because of the AI. To understand why, look at a company's CRM. That's where the skeletons are. I've spent the last few months drilling into how our customers actually organize their data. One tracks customer health in a deal field. Another tracks it through projects. Another buried it inside a custom object their ops team built two years ago and forgot to document. Every schema is a fingerprint. No two are the same. The engineering instinct is to standardize. Pick the one true way, build your product around it, ship it. That works until you meet your second customer. Most AI companies skip past this part. They want to get to the fun stuff. The agents, the models, the prompts. They treat the data layer like a tax you pay before you get to the real product. I've learned the data layer is the product. We started Cora focused on post-sales. Most of the revenue lives there, but no one's been able to crack it because the data is so messy. At times, it felt like a fool's errand. Pre-sales data is relatively clean. A lead is a lead. An email is an email. Post-sales data is a record of a relationship, and every company records relationships differently. I used to think the hard part of building an AI product was the AI. Turns out the hard part is the layer between the customer's data and the model. Solve that, and everything else follows.
-
Jesal Gadhia reacted on thisMany learnings from building our app alongside such talented teams, and learning what huge CS teams need Agents for! It's amazing what our small team can build for so many companies :)Jesal Gadhia reacted on thisWe crossed $3B in revenue managed by Cora agents just months after launch... so we hopped on a plane to hear from customers on the outcomes Cora is driving. Some of the highlights we heard: → HeyGen reactivated hundreds of dormant customers, with direct impact to net dollar retention → BetterUp saw 5,000 hours and at least a million dollars saved from automating manual work → Flock re-engaged thousands of long-tail customers with personalized touch
-
Jesal Gadhia reacted on thisIf you are a CCO, CRO or RevOps leader and still running post-sales on Gainsight, Vitally, Planhat, Catalyst or Salesforce, please let me know and I'll connect you to Amir Ghowsi!Jesal Gadhia reacted on thisWe crossed $3B in revenue managed by Cora agents just months after launch... so we hopped on a plane to hear from customers on the outcomes Cora is driving. Some of the highlights we heard: → HeyGen reactivated hundreds of dormant customers, with direct impact to net dollar retention → BetterUp saw 5,000 hours and at least a million dollars saved from automating manual work → Flock re-engaged thousands of long-tail customers with personalized touch
-
Jesal Gadhia reacted on thisJesal Gadhia reacted on thisI'm excited to share I'm starting a new role as Director of Partnerships at Syndio. The people have always been the best part of work for me, and I'm grateful to join some familiar faces and incredible leaders, including Erik Darby, Shonna Waters, PhD, and Erin McClintock. Syndio has a real opportunity to grow through partners, and I can’t wait to help more organizations make payroll, their biggest spend, work the way they intend. Leaving BetterUp after 5.5 years was a tough decision. I’m proud of the partnership ecosystem we built and am lucky to call so many former colleagues friends. Special thanks to Thomas Van Gilder, MD, JD, MPH, Erik Darby, Shevaun (Bequette) Lee, Mike Hodel, and Chitra Subramaniam for being amazing leaders to work for and learn from.
Experience & Education
-
cora.ai
********** * ***
-
********** ***** *********** *********
**** undefined undefined
-
View Jesal’s full experience
By clicking Continue to join or sign in, you agree to LinkedIn’s User Agreement, Privacy Policy, and Cookie Policy.
Welcome back
By clicking Continue to join or sign in, you agree to LinkedIn’s User Agreement, Privacy Policy, and Cookie Policy.
New to LinkedIn? Join now
Recommendations received
7 people have recommended Jesal
Join now to viewView Jesal’s full profile
-
See who you know in common
-
Get introduced
-
Contact Jesal directly
Other similar profiles
-
Victoria Alden
Victoria Alden
Wilshire Group Financial Services
4K followersLos Angeles Metropolitan Area
Explore more posts
-
Mayur Jain
GoDaddy • 4K followers
GoDaddy Airo Builder isn’t just “describe it and ship it.” Under the hood there’s a full set of settings so you can run and grow your project properly. In this walkthrough we show where everything lives: → Access & delegation who can work on the project → Backups automatic daily backups and version history so you can restore anytime → Analytics built-in dashboard for page views, visits (no extra tools) → Database managed MySQL you can view and manage right in the UI → Secrets secure storage for API keys and credentials, encrypted and separate from your code Try it here: https://lnkd.in/gNGjZHwV #GoDaddy #AiroBuilder #WebDevelopment #NoCode #SmallBusiness
7
-
Adolfo Usier
Open Crabs • 4K followers
Vibe coding tools are making app prototyping much faster for product teams. They let designers and non technical makers turn ideas into working pages quickly. That speed is great for testing ideas, but it also means teams must add review and testing steps. What matters is pairing vibe builds with good code review, security checks, and monitoring. Check the company post for the full breakdown and examples. #meetneuraai #neuraai #vibecoding #aicode #webdev #product.
1
-
Robbie Abed
ProducerSpark • 35K followers
Claude is the computer. If you're building software, your product has to work in Claude to grow. But you can only really win in this space if someone types a prompt in Claude and gets an outcome they can only get when your plugin is installed. That's where I think software engineering is headed.
6
-
Lucas Tieleman
Optiversal • 3K followers
We shipped 3 big product updates in the last few. Been so heads down building I haven't had time to post about it. Quick rundown: 1. WYSIWYG Page Editor Your content team can now control page layout without bugging engineering. Add product modules, FAQs, carousels, custom assets. No dev tickets, no waiting. So if you've got a landing page driving 10K visits but converting below average, you can add a new CTA and test it by end of day. Not three sprints from now. 2. Top Movers Dashboard that shows which pages are accelerating, which are tanking, and which are driving the most right now. You can toggle between views, impressions, conversions. Weekly or monthly. Say your "outdoor patio furniture" category is picking up steam heading into spring. You catch it early, enrich the content, and ride the wave instead of reacting after the peak passes. 3. Performance Views Traffic, revenue, conversions, AOV, impressions in one place. Page-level and program-level. Plus automated summaries so you're not squinting at charts trying to figure out what changed. You pull up a page, see revenue up 30% but traffic flat. That tells you conversion improved. Now you know what to replicate. All three point at the same thing: shrinking the gap between seeing a problem and actually fixing it. More coming soon. Let's go Optiversal.
30
-
Fahim ul Haq
Fenzo • 26K followers
There is no better combination of system design and product experience than what Spotify pulls off every December. Every year I look at it and think: "This is what great engineering feels like." If you're a builder, ask yourself: 👉 Are you making something that scales and sticks emotionally? 👉 Do your users feel seen, or just processed? Think about what they pull off: - 600M+ users - Insane volumes of data stitched together across edge devices, stream processors, batch jobs - A global rollout that doesn’t break — and delivers a deeply personal experience to everyone at once - That’s neither easy nor is it about good infrastructure. It all hinges on how they orchestrate it. Wrapped works because it turns raw data into identity. It’s such a potent viral loop that giants like Apple Music and YouTube had to adopt the format just to stay part of the conversation. Behind the scenes, it relies on cutting-edge distributed systems to crunch data, but on the surface, that heavy lifting is distilled into a moment of pure, personalized delight. Anyway, I’m off to replay my 2025 case files again. What stood out to you in your Wrapped this year — as a user or as a builder? #SpotifyWrapped #SystemDesign #ProductThinking #EngineeringLeadership #DeveloperExperience
19
2 Comments -
Dominick Luna
Capxel • 1K followers
A builder question worth carrying through the weekend: If we made this process ten times faster, would we be proud of the result? If the answer is no, automation is not the first problem. The process may be unclear. The incentives may be wrong. The decision may not have an owner. Speed amplifies structure. It does not repair it. Before we add another layer of technology, make sure the work is pointed somewhere worth accelerating. #Founders #Technology #Execution
-
Jonathan Katz
Matcha Consulting • 2K followers
One of the biggest mistakes I see builders make is using a direct linked record relationship when what they actually need is a junction table. It's easy to miss when you need one, but understanding when and how to use junction tables will give you a whole new way to model data, build automations, and create user flows in interfaces. Luckily, it's not as complicated as it sounds! Check out today's JK Theory for more. Link in comments.
9
5 Comments
Explore top content on LinkedIn
Find curated posts and insights for relevant topics all in one place.
View top content