1,000+ Claude Prompts Top Professionals Actually Use at Work
Claude can be your analyst, editor, and strategist.
But most professionals are using it to fix grammar.
These 1,000+ Claude prompts take it from grammar tool to your most powerful AI work assistant.
Sign up for Superhuman AI and get:
1,000+ ready-to-use Claude prompts to get real work done in minutes — researched, tested, and used by professionals at Google, Microsoft, and NASA
Superhuman AI newsletter (4 min daily) so you keep learning new AI tools and skills to stay ahead in your career — the prompts are just the beginning
Most AI tools start with the model. GoMarble started with a year of doing the work by hand.
Aditya Sriram is one of three co-founders of GoMarble, an AI agent platform for paid media teams at brands and agencies, based in Bengaluru. He leads growth and operations alongside CEO Palkush Rai Chawla and CTO Swapnil Bari.
Starter Story's page lists the company at $50,000 a month with 3 founders and 10 employees, though the article itself does not state revenue. One number is easy to misread: the more than $250M in monthly ad spend connected to GoMarble is spend it manages for clients, not money it earns. Here is how the founders got there.
A Year of Running Ads by Hand
The three founders came with strong backgrounds. Palkush had built two companies and worked on monetization at ShareChat and ITC. Swapnil was a staff engineer at ShareChat and Goldman Sachs. Aditya led growth at Rupeek after working at BCG and McKinsey, and he and Palkush first met as first-year students at IIT Kharagpur in 2007.
From May 2024 to May 2025 they ran Google and Meta ads for 20 US brands, about $200,000 a month in spend. That agency-style work also brought in their early revenue, before the product existed.
Much of their time went to one tedious job: assembling each account's history by hand from ad dashboards, exports, and spreadsheets. ChatGPT could not help much, because it only knew what they pasted into it.
The Problem Was Never the Model
That frustration gave them their core insight. In Aditya's words, "the hard part wasn't the AI, it was the context."
So they built an internal tool that connected the data, kept each brand's history, and ran the investigations they repeated every week. That tool became GoMarble.
In April 2025 they released one of the early advertising MCP servers, which let marketers query Meta Ads from Claude, and followed it with a Google Ads version. The full product launched in October 2025.
What the Agents Actually Do
GoMarble connects Meta, Google, TikTok, LinkedIn, and Microsoft Ads, plus GA4 and Shopify. Its agents monitor campaigns, explain why performance changed, analyze creative, and propose actions such as budget shifts, bid and status edits, negative keywords, and new ad launches.
The platform is now connected to more than 12,000 ad accounts. The article does not list pricing or name the tech stack.
Trust Before Autonomy
Changing live campaigns puts client money at risk, which made trust the hardest problem to solve. Aditya puts it bluntly: "If AI touches real money, 'trust me' is not a safety system."
The product is built around that idea. Agents act only within permissions the team sets, every action is logged, and they keep each brand's context so their advice improves over time. They show their work, and they gain more autonomy only after proving they are reliable.

Launches, Creators, and 300 Pages
Growth came from several channels working together:
Founder-led LinkedIn content. Three major launches focused on new capabilities. The MCP launch post drew more than 5,000 comments, and launch posts passed 500,000 impressions in their first 12 hours.
Creator partnerships. More than 100 paid media creators across LinkedIn, X, Instagram, and TikTok.
A place inside AI assistants. The product works inside Claude and ChatGPT and is listed in their integration directories.
Programmatic SEO. More than 300 pages, including ad cost research, an ad library, tool comparisons, and AI workflows. AI agents help plan, research, build, and review them under rules people approve.
Community and launches. A Slack community of more than 1,000 marketers, and a #1 Product of the Day finish on Product Hunt.
The Webinar That Brought Zero Signups
Not everything worked. A webinar drew 62 registrants, nearly half of them existing users, and produced zero new signups.
They also cut their own signup numbers on purpose. In April 2026 they stopped accepting personal email signups, trading a smaller signup count for more real teams. Their retention data points the same way: teams that, within 30 days, connected both Google and Meta, had 10 or more ad accounts, added a second user, and used GoMarble from an AI assistant were more than twice as likely to still be customers at day 120.
That finding shaped one of Aditya's rules: "judge every channel on paying customers, not signups."
What Founders Can Take From It
The lessons Aditya draws apply well beyond ad tech:
Make money doing the work first. Running ads for clients gave them validation, real data, and a list of problems worth automating.
Earn trust before asking for it. When software acts on someone's money, permissions, logs, and visible reasoning matter as much as the result.
Protect your signal. Fewer, better signups made it easier to see what real teams needed.
Find the habits that predict retention. Then measure every channel against paying customers, not top-of-funnel numbers.
Build an audience before launch. Giving people something useful early, like the MCP servers, built an audience before the full product launched.
The article does not lay out a formal roadmap. You can see the product at gomarble.ai.
Credit and sources: this edition is based on the Starter Story interview with Aditya Sriram of GoMarble.

