🎉 Customer story: How Intelswift shipped AI agents in 1 month with 782% ROI
Read the story
Pinkesh Shah has experienced a career path that could fill two lifetimes: he began as a software engineer, rose to Vice President of Product Management at a NASDAQ‑listed company, and then founded his own startups. Driven by passion, an insatiable hunger to go from zero to one repeatedly, and a deep motivation to solve customers’ problems, he has invaluable insights to share.
What are the key lessons for young product professionals who want to make the most of their careers? How does one transition from software engineer to VP of Product at a global enterprise?
Find out in the latest episode of PM Talks.
‍
It’s been a fascinating journey. As an engineer who accidentally became a product manager many years ago, I still remember the day our CEO approached me at a small startup in Houston. He told me they wanted to overhaul the compensation model for all three engineering managers—shifting our bonus from overall company performance to the success of our individual products.
Accepting that change completely altered my career path. Although I kept the title of engineering manager and continued running the same team and product, my focus shifted dramatically. Instead of obsessing over code refactors or architectural purity, I began asking, “Who is the customer, and how can we truly delight them?” After all, my bonus now depended solely on product revenue—if customers weren’t happy, they wouldn’t buy more, and I wouldn’t earn my bonus. That single decision became my unofficial introduction to product management.
Since then, I’ve led multiple product management, design, and marketing teams. Now, I’m returning to my startup roots to launch DelightLoop, an AI‑agent software company. Our mission is to leverage AI to foster human connections and help marketers generate predictable leads.
It’s hard to pin down a single formula for success, but there are a few common catalysts. Early on, I became obsessed with solving real customer problems—digging beneath surface complaints, spending time in users’ shoes, and asking what their day‑to‑day looks like. Only by understanding the true pain points can you begin to craft solutions that matter.
From there, it’s all about a growth‑mindset approach to experimentation. What many call “hypothesis‑driven design” is really just a cycle of testing different ways to solve an age‑old problem until you land on something fresh and delightful. That willingness to iterate relentlessly is what helped me excel in every role—and, honestly, what got me noticed enough to climb the so‑called corporate ladder.
‍
Now, as a hiring manager, I look for three key lenses when considering someone for a senior product role. First are functional skills: mastering the art and science of every step—discovery, prototyping, validation, pricing, packaging, messaging, and branding. Second are leadership skills: storytelling and the ability to influence people who don’t report to you, making cross‑functional collaboration feel seamless. Third are domain skills: a genuine passion for your industry and deep empathy for your user persona, staying attuned to trends and real‑world workflows. Excel in all three, and you’re ready for that next promotion.
Balancing career logic with personal passion has been a guiding tension for me. When I moved into the startup world, it wasn’t because the business model was rock‑solid—it was because I was driven by passion for the work itself.
While setting up my product team in India, I sifted through hundreds of resumes from top B‑school graduates and still couldn’t find a single qualified product manager. Money wasn’t the issue; we could pay a lot at McAfee. So I promoted two of our best engineers in Bangalore into PM roles and invested in their training through U.S. programs.
On the 20‑hour flight home, it hit me: India has one of the largest pools of engineers and IT professionals in the world, yet product management discipline is almost nonexistent. That moment sparked the idea for the Institute of Product Leadership—arguably the world’s first business school for product people.
Today, the Institute operates across multiple campuses, and nearly 6,000 graduates have completed programs ranging from short courses to a fully accredited Executive MBA in Product Leadership. Our mission is to give engineers, analysts, project managers, and marketers a clear, systematic path into skilled product management.
I love the thrill of going from zero to one—building new capabilities and defining categories—far more than optimizing established machinery.
There are many flavors of product people. Some thrive scaling from $100 million to $500 million; others grow a billion‑dollar business even larger. All are product management, but the mindset and tactics differ.
My passion has always been rapid innovation and creating something entirely new. When a role opened in a cybersecurity company—a domain I’ve long been obsessed with—it felt like the perfect next challenge.
By then, the Institute of Product Leadership was thriving under its new CEO, an ex‑Oracle product manager. That let me return to my cybersecurity roots at Qualys, tackling cutting‑edge problems in demand generation and risk quantification—areas I believe remain ripe for breakthrough solutions.
I’ve always been driven by a passion for solving customer problems—and at Qualys, that instinct surfaced again. Marketing costs have skyrocketed, especially in B2B, making customer acquisition both expensive and unpredictable. Despite an ecosystem of 300+ tools for automation and intelligence, everything felt fragmented—no single platform tied it all together.
That inspired DelightLoop’s core mission: deliver a systems‑driven automation platform that makes lead generation, deal‑closing, and churn reduction predictable. We boiled it down to three outcomes: generate high‑quality, high‑intent leads; shorten the sales cycle; and minimize renewal churn—for large companies, even a 1% improvement can mean millions saved.
To do this, we’re building autonomous “co‑pilots” for every marketing specialty—event management, performance marketing, ops, and more—powered by AI agents. Our first agent, Gifty, tackles hyper‑personalized gifting: sending the right gift to the right prospect at the right moment. In trials, Gifty drove a 30% lift in response rates versus purely digital outreach.
Digital channels are increasingly noisy, and thoughtful physical gifts cut through the clutter to forge genuine connections. At scale, DelightLoop aims to weave delight into every touchpoint—helping brands build human relationships through smart, AI‑driven assistance. That continuous loop of delight is what our platform is all about.
The core skill sets don’t change whether you’re a two‑person startup or leading a hundred‑strong product team at a multi‑billion‑dollar company. What shifts is where you spend your time—and that depends on your company’s stage.
In a startup, you live in “hypothesis design.” You rapidly brainstorm experiments, decide which few are worth pursuing, and iterate at speed. One wrong turn can cost months—or the whole venture—so prioritization and a growth‑mindset are crucial.
As you gather early customers, you move into “customer design.” You co‑create with pilot users, say no to misaligned segments, and refine your product vision around real feedback. This phase is all about selecting the right niches and becoming indispensable to them.
Once you have scale, “product design” takes center stage. You sweat the user experience—handling edge cases, tailoring flows for both small and large clients, and polishing every interaction. It’s not just UX; it’s the end‑to‑end customer experience.
Finally, in mature organizations, you shift into “revenue design.” You manage product portfolios, explore adjacent markets, model pricing, and build upsell and cross‑sell strategies. Here, your focus is on sustainable growth and unlocking new revenue streams.
No matter the company size, mastering each design phase—hypothesis, customer, product, and revenue—is the framework for advancing your product career.
‍
People often confuse systems design with structure—and structure with process. Even in a sub‑$10 million‑revenue startup, you need solid structure: clear frameworks for decision‑making, handoffs, and priorities. Chaos doesn’t mean “no rules.” A product manager shouldn’t wake up each day to pure improvisation—that’s a recipe for costly mistakes.
As you scale toward thousands of customers, formal processes become just as critical. One unvetted UX tweak or pricing change can ripple through a large user base, causing real damage (like the recent cybersecurity glitch that grounded airlines). In a ten‑customer pilot you can personally check every user, but at 5,000 customers you need automated checkpoints, reviews, and gates. The trick is balancing minimal process (to keep hypotheses flowing fast) with enough structure to protect your growing customer ecosystem.
.avif)
The pace of AI development makes any prediction beyond nine months a fool’s errand. Yet for product managers, this AI‑first era is thrilling—every week brings new capabilities and possibilities.
Many worry AI will steal their jobs over the next two to three years. It’s true some roles will be automated, but the real risk isn’t AI taking jobs—it’s you not learning AI. Those who master AI tools will unlock better growth, faster promotions, and more strategic roles.
Experimentation speed has exploded—from three months to three days to three hours. Tools like v0, Bolt, and Cursor can spin up full interfaces from just a vision and user insight. No more waiting on UX designers or mastering Figma: you can prototype in hours and test with real users the next day.
The research process itself is being revolutionized. Competitive analyses, user interviews, even strategic roadmaps can be generated and refined through AI prompts. Prompt engineering has become the new superpower—often more impactful than learning Python. Building fluency in AI workflows will redefine how product managers work in the coming year.
Routine tasks—writing user stories, tracking project progress, basic backlog grooming—are prime candidates for automation. If your role today is purely about translation or coordination, AI will displace you. But for those ready to focus on decision‑making—senior PMs and directors—AI frees you to tackle higher‑value challenges.
Finally, PMs must rethink explainable AI and business models. Delivering intuitive, transparent AI (XAI) is table stakes, but the bigger shift is monetizing outcomes over seats. Imagine RAAS—Results‑as‑a‑Service—or OAS—Outcome‑as‑a‑Service—where customers pay for delivered value, not monthly licenses. Embracing these changes will open unprecedented opportunities for product leaders.
Venture capital firms will need time to adjust to outcome‑based models—it breaks the predictable “hockey stick” revenue curve they love. But from the customer’s perspective, it’s a no‑brainer. Take our webinar use case: a marketing manager needs 40 more qualified registrants, so our Gifty agent identifies the ICP, nurtures them digitally, then sends each prospect a hyper‑personalized gift. The result? A 30% uptick in attendees.
Customers pay only for the gifts delivered—and if we hit that 30% lift, we earn an outcome‑based bonus. Our skin is in the game: the product must perform, or we don’t get paid. This model lets us charge more per customer, yet they happily accept it because they’re seeing real results. I expect many vendors to shift to credit‑, usage‑, or results‑based pricing in the coming months—aligning value with cost and driving true customer delight.
.avif)
It’s a tough balance: as a founder, you can’t afford to embrace failure; as a practitioner, you know setbacks are inevitable. What matters is how quickly you return to hypothesis design—treat every stumble as fuel for your next experiment.
I see two big shifts coming in marketing. First, AI will automate roughly 80% of the tactical work across demand gen, brand, events, performance, and ops. Instead of ten layers of specialists, you’ll need just one skilled “GTM engineer” who orchestrates AI agents to run a team of ten.
Second, the current patchwork of dozens of tools will consolidate into a handful of autonomous marketing platforms. No more stitching together 48 apps via Zapier—future CMOs will rely on end‑to‑end systems that handle everything from lead capture to nurture without becoming an IT department.
The payoff is huge: program spend and headcount drop, acquisition speed rises, and cost per customer plummets. That’s why I’m extremely bullish on the next few quarters—and years—ahead for marketing powered by AI.
Finding people who are truly passionate about your mission is the biggest challenge for any startup. We see plenty of resumes, but what matters is genuine excitement for solving our unique problem—using AI to create delightful marketing experiences.
I think “delivered delight”—and I mean quantified delight. So when your customer is so thrilled (not just satisfied or happy) that they become unsolicited evangelists, referring ten more people and acting as your open recommendation engine. To me, that’s the North Star metric every product manager should aim for—revenue is merely the byproduct of a genuinely delighted customer base.
I think digital tools—FunnelStory, Pendo, and the like—offer great visibility into how people use your product. But whenever you can, especially early in your career, carve out a dedicated day each week for direct customer conversations. For me, that’s Wednesday: I spend the whole day talking with users about how they’re implementing the product, the challenges they face, and the novel ways they’ve repurposed features we never intended.
Those human insights are invaluable. In enterprise settings, it’s common to build for one scenario only to learn customers are using a feature in a completely different way. No matter how busy you get, don’t outsource or delegate this work—personal customer interviews are essential to understanding the real user journey.
‍
.avif)
There’s so much you can buy or integrate via out‑of‑the‑box connectors that you rarely need to build from scratch anymore.
That’s why I always recommend focusing on your core value proposition and partnering or purchasing everything else. By buying what you can and only building what differentiates you, you get to market faster and free up your team to innovate where it really counts.
‍
The agent we have in production today is called Gifty. Gifty does three core things: it finds the right gift for the right person, sends it at the perfect moment, and orchestrates the entire process—from selection and discovery to physical delivery and tracking—ensuring your recipient takes the intended action as soon as the box arrives.
To make Gifty truly effective, we integrate with multiple signal sources to pinpoint the ideal buyer. Take our webinar example: a demand‑generation leader running 15–20 webinars a year often exhausts her known contacts and needs fresh, high‑intent prospects. By tapping into CRM data, job‑change alerts, and competitor reviews on sites like G2 or Gartner, Gifty zeroes in on people most likely to convert.
We call this insight layer DelightSense. It aggregates signals from your CRM and external marketplaces, prioritizes prospects based on buying intent, and triggers the optimal gifting workflow. Once you’ve chosen who to target, the next module—DelightEngage—takes over, automating hyper‑personalized, omni‑channel outreach via email, LinkedIn, WhatsApp, Slack, and more until you capture the prospect’s attention.
These modules power Gifty’s seamless gifting agent, giving marketing managers a turnkey way to build genuine connections with key prospects. Looking ahead, our next AI teammate is Yvette. Like Gifty for gifting, Yvette will plan and execute entire events—webinars, product launches, and more—within seconds. You’ll simply tell Yvette your budget and attendance goals, and she’ll generate a full plan (narrative, speakers, landing pages, event portal setup) and then execute it behind the scenes.
By treating AI agents as teammates—whether for gifting or event management—DelightLoop enables marketing leaders to operate with startup‑speed, enterprise‑scale predictability, and a continuous loop of customer delight.
From Pinkesh’s inspiring story to practical tips on climbing the corporate ladder, we covered a lot in today’s episode! We hope you learned something—and don’t forget…
You can contribute to the topics we prepare for our PM talks. Post your desired question or topic below, and we’ll happily consider it for future episodes!
You’ve just read an interview from our podcast, where we speak with product leaders who share their experiences. Follow us on Spotify or YouTube for more episodes.
.jpg)
.png)
.png){kind=logo}
