Selling Go-To-Market Agents to Agencies, Funded by Running the Agency Work First
Santanu Dasgupta took an AI orchestration platform for lead generation and campaign follow-up to $3K MRR in four weeks by selling into agencies they already knew, with services revenue paying for the software build
Continue exploring this workflow
The Strategy
Santanu Dasgupta spent twenty years in go-to-market before building Meerkats.ai, and that ordering matters. The platform runs the work an SDR or a small agency team would otherwise do by hand: capturing and enriching relationship data, generating leads, running campaigns and following up, all driven from a chat interface rather than a dashboard of settings. The commercial structure is the more interesting part. They funded the software build with agency services revenue — doing the work manually for clients while turning the repeatable parts into product. That is the productised-services path several operators in this directory have taken, and it has a specific advantage: the services work tells you which parts of the job are actually worth automating, rather than guessing from the outside. The platform is built on Supabase and Fly.io, and stays deliberately model-agnostic, letting customers choose between Claude, OpenAI and others rather than locking to one provider. They position that flexibility directly against the first-party agent platforms, which is a reasonable wedge when selling to agencies that already have opinions about which model they trust. Distribution came from the twenty years, not from launch tactics: cold outbound to agencies, warm introductions, and LinkedIn engagement, reaching $3K MRR in four weeks. They also took non-dilutive support — a University of Chicago Polsky Center grant plus cloud and model credits — which is the quiet way a lot of these early agent products actually get funded.
How It Works
Start from a domain you have worked in for years, so you know which parts of the job are repetitive and which need judgment.
Sell the manual service first and let that revenue fund the product build, rather than building a platform speculatively.
Use the services engagements to identify which steps are genuinely worth automating before writing them into product.
Automate the full go-to-market loop — capturing and enriching relationship data, generating leads, running campaigns, following up — rather than one slice of it.
Drive the whole thing from a chat interface so operators describe outcomes instead of configuring workflows.
Stay model-agnostic so customers can choose Claude, OpenAI or another provider, and use that flexibility as a wedge against locked-in platforms.
Sell into agencies through cold outbound and warm introductions from your existing network rather than a public launch.
Take non-dilutive funding where available — grants and cloud or model credits — to extend the runway during the build.
Results
Santanu Dasgupta reports reaching $3K MRR within four weeks of a May 2026 launch, with the build funded by agency services revenue plus a University of Chicago Polsky Center grant and credits from Azure, OpenAI and Anthropic. Customer acquisition came from cold outbound to agencies, warm network introductions and LinkedIn engagement. Figures are self-reported and have not been independently verified.
Our Take
The pattern here is the one worth extracting: run the service manually, get paid for it, and let the client work tell you what the product should be. Almost every failed agent product we come across was built the other way round. Selling into agencies is also a smart wedge, because agencies buy tools that let them take on more clients without more headcount, which is a far easier sale than convincing an end business to change how it works. $3K MRR in four weeks is a real but early number, and it rests on twenty years of network — anyone reading this without that network should expect the distribution half to take considerably longer than the build. Best for operators with deep experience in a specific service who are considering turning it into an agent product.
Frequently Asked Questions
The practical questions a builder or operator is likely to ask before trying a strategy like this.
What does this marketing agencies AI sales agent actually do?
This marketing agencies AI sales agent is a real workflow where the agent takes on an operational job, not just a brainstorming task. Selling Go-To-Market Agents to Agencies, Funded by Running the Agency Work First shows what that looks like in practice. Santanu Dasgupta took an AI orchestration platform for lead generation and campaign follow-up to $3K MRR in four weeks by selling into agencies they already knew, with services revenue paying for the software build The practical value comes from the agent handling repeatable business work with enough autonomy that a human only steps in after context has already been gathered.
Who should use a marketing agencies AI sales agent like this?
This example is most relevant for marketing agencies operators. It is especially relevant for businesses where speed to lead, after-hours coverage, or consistent intake quality directly affects revenue. The category here is Marketing & Sales, which means the best fit is a team looking to turn a manual bottleneck into a repeatable system with a marketing agencies AI sales agent.
Which tools are used in this marketing agencies AI sales agent setup?
The source names Supabase, Fly.io, Claude, OpenAI. That matters because one of the strongest signals in this directory is whether the operator shared the actual stack. Named tools make a marketing agencies AI sales agent strategy far more useful than vague claims about “an AI system” doing the work.
How hard is it to implement a marketing agencies AI sales agent like this?
Advanced difficulty is the current read. The listing suggests a launch window of weeks. Startup cost is listed as $200+/mo. We were able to extract 8 concrete workflow steps from the source. We would treat a marketing agencies AI sales agent like this as a workflow that needs real business context, testing, and exception handling rather than something you should copy blindly from one prompt.
What results can a marketing agencies AI sales agent produce?
Santanu Dasgupta reports reaching $3K MRR within four weeks of a May 2026 launch, with the build funded by agency services revenue plus a University of Chicago Polsky Center grant and credits from Azure, OpenAI and Anthropic. Customer acquisition came from cold outbound to agencies, warm network introductions and LinkedIn engagement. Figures are self-reported and have not been independently verified.
How credible is this marketing agencies AI sales agent case study?
Right now the evidence comes from an Indie Hackers post. That is enough for us to study and curate the workflow, but not enough on its own to treat this marketing agencies AI sales agent like an audited case study. We look for named tools, concrete results, and enough workflow detail to understand what was actually deployed, then we add our own editorial judgment on top.
Related Strategies
More AI agent strategies you might find useful
Fully Autonomous Meta Ads Manager Built on OpenClaw
A fully autonomous Meta ads operation running on OpenClaw for $0 per month that …
SaaStr Replaced a 10 Person Sales Team With 20 AI Agents and 1.2 Humans
A sales team of 10 SDRs and AEs was replaced by 20 AI agents managed by 1.2 huma…
Running a Company Entirely With AI Agents: What Worked and What Broke
Seven AI agents ran a company on a single VPS for a week. They pivoted the busin…
Building a $3,500 AI Customer Service Chatbot Live From Scratch
A full live build of a customer service chatbot with knowledge base retrieval, l…
12 n8n Workflows That Replaced $3,000 Per Month in Manual Work
Twelve n8n automations eliminated $3,000 per month in manual tasks covering lead…
An AI Marketing Product That Grew to $30K MRR by Dominating Reddit
An AI marketing tool reached $30K MRR by building distribution entirely on Reddi…
Want more strategies like this?
Get weekly AI agent case studies in your inbox.