Shalin Jain built HappyFox to $20M+ ARR, bootstrapped, with about 100 employees. He went from 5 support reps at $1M ARR to 6 support reps at $20M ARR. Same story on sales: 6 sellers at $3M, still 6 sellers at $20M. He recently shared how his team did it by treating support data as their primary GTM engine. Here are the 5 most actionable takeaways.
1. Your Support Queue Already Has the Expansion Signals. You Just Need Something Reading Them.
Most B2B companies treat expansion as a sales-led or CS-led motion. Shalin found the highest-ROI signals were already sitting in support tickets.
Customers mention needing more seats. They ask about features on a higher plan. They reference products they don’t have yet. Support reps see this every day, but they’re not incentivized (or trained) to route it to sales. And sales doesn’t have access to the ticket queue.
HappyFox built an AI agent that scans every incoming ticket for expansion language. Out of 5,000 tickets a month, it might flag 13 active opportunities. But those 13 are high-context, high-intent, and came from the person actually using the product. Not from a lead score. Not from a product-qualified signal that may or may not mean anything. From a customer who basically said “I want more.”
The cost to run that agent: about $8 to $10 total.
Do this now: Even without AI, create a Slack channel where support can flag any ticket that mentions adding seats, upgrading, or asking about features the customer doesn’t have. You’ll be surprised how much pipeline is hiding in plain sight.
2. Bridge CRM Data Into Support (and Support Data Into Your CRM). Both Directions Matter.
This was one of the sharpest tactical moves in Shalin’s playbook. He built two agents that work in opposite directions.
Direction one: CRM to support. When a trial customer or active deal submits a support ticket, the agent pulls deal stage, MRR, and context from the CRM and surfaces it to the support rep. So support knows this isn’t a random free-tier user. It’s a $50K deal in evaluation. That ticket gets priority treatment, not the standard queue.
Direction two: support to CRM. When a customer has open tickets about a broken integration or a frustrating bug, the sales rep gets alerted before their next scheduled call. Nothing kills a deal faster than a rep pushing on an order form while the customer is struggling with something nobody told them about.
Most companies do neither of these. The few that do typically only go one direction. Doing both eliminates the worst coordination failures between sales and support.
Do this now: At minimum, tag every support ticket from a customer with an open deal in your CRM. Even a simple flag that says “active opportunity” changes how support prioritizes.
3. Churn Signals Show Up in Support Conversations Long Before They Show Up in Usage Data.
Product usage drops are a lagging indicator of churn. Support conversations are a leading one.
When a company gets acquired, the new leadership often writes into support saying “I’m re-evaluating our vendor stack.” When a champion leaves and a new stakeholder takes over, they’ll submit a ticket saying “I need an overview of what we’re paying for.” These are turning points in a customer relationship, even one that’s been healthy for 8 years.
Shalin built a churn signal agent that flags this language automatically. The key categories: leadership change mentions, acquisition or merger references, “re-evaluating” or “comparing alternatives” language, and requests for billing or contract summaries from new contacts.
Do this now: Search your support tickets from the last 6 months for the words “re-evaluate,” “new here,” “took over,” and “comparing.” Count how many of those accounts later churned. That’s your baseline for how much revenue you’re losing by not catching these signals.
4. Your Marketing Team’s Best Content Is Hiding in Support Threads.
This was the one Shalin said his marketing team loved the most. He built two agents: a testimonial identifier and a case study finder.
The testimonial agent flags any moment where a customer expresses genuine enthusiasm in a support conversation. Someone writes “Scheduled tickets completely changed how we run our entire IT operation,” and that quote gets surfaced to marketing for the relevant feature landing page. Real language from a real customer, not something a copywriter guessed at.
The case study agent flags conversations where customers describe how they implemented something, what changed in their workflow, or what results they got. Even if it was the success team or sales team that did the actual work, the mention in a support thread is enough to surface the opportunity.
Most marketing teams build case studies on a quarterly cycle, chasing customers who may or may not want to participate. This approach catches the moment of enthusiasm when it actually happens.
Do this now: Ask your support lead to forward you the 5 most positive customer interactions from the last month. You’ll probably find at least 2 usable quotes and 1 case study candidate.
5. Start With the “Cart Add-On” Play. It’s the Easiest Win.
One of Shalin’s simplest agents tracked every time a customer requested additional seats or licenses. Instead of just processing the request, it triggered a sales touch to explore what else the customer might need.
The logic is the same as e-commerce: if someone’s already adding something to the cart, that’s the moment to suggest adjacent products they haven’t considered. A customer who wants 2 more seats probably also needs the workflow automation module they’ve been ignoring. Or the reporting tier they outgrew 6 months ago.
This is the lowest-effort, highest-return version of expansion selling. The customer already signaled buying intent. You’re just asking “what else?” at exactly the right time.
Do this now: Look at your last 20 seat-addition or plan-upgrade requests. In how many of those cases did someone also talk to the customer about adjacent products? If the answer is close to zero, you’re leaving money on the table every week.
Shalin ran all 8 of his support-to-GTM agents for under $100 total, on first-party data, with same-day results. The details of his full agent setup are worth watching if you want the implementation specifics. But the core insight is simple: your customers are already telling you what they need, what’s frustrating them, and what they love. It’s all in your support queue. The question is whether anyone besides your support team is actually listening.
Want to meet Shalin and 1000+ other B2B + AI leaders sharing real playbooks like this? Join us at SaaStr AI Annual 2026, May 12-14 in SF Bay.
