“Build Cars, Not Faster Horses: What It Takes to Win in the Agentic AI Era” | The Data Huddle – Erez Shilon, Senior Director of Product Management at Gong

By Dana Bilder Morozova
July 15, 2026 | 5 min read

The Data Huddle by Yuki is a podcast about modern data leaders at global companies. The people who make data work every day, but rarely get the spotlight.

Honest conversations about their careers, wins, mistakes, and how they see the future of AI and data. What does it actually take to build a product org that keeps winning, year after year, while everyone else is chasing the same AI wave?

In this episode, Dana Bilder-Morozova sits down with Erez Shilon, Senior Director of Product Management at Gong, for a candid conversation about product strategy in the agentic AI era. Not the hype version. The version where data foundations, customer obsession, and a strong point of view actually decide who wins.

On This Episode

Anyone can open Claude or ChatGPT today and prototype “their own Gong” in a weekend. So what actually protects a category leader?

Erez’s answer isn’t a feature. It’s a set of disciplines: falling in love with the problem instead of the technology, ingesting and modeling data the right way, keeping product teams obsessively close to customers, and having the courage to build what customers need instead of what they ask for.

They covered a lot of ground: how Erez went from wanting to be a construction project manager to leading product at an Israeli unicorn, why “faster horses” is the most dangerous thing a PM can ship, what actually happens inside Gong when a new AI trend shows up, and why nobody ever asks a product manager to show up to work.

About the Guest

Erez Shilon is Senior Director of Product Management at Gong, where he leads the data side of the product organization. His relationship with Gong actually predates his employment there. Back in 2016, while working at a small startup that would later become Own Backup, he shared a WeWork space in Herzliya with Gong’s founders, Amit Bendor and Eilon Reshef, before Gong was the company it is today. He started following them, and when the opportunity to interview at Gong came up more than four years ago, he didn’t think twice.

Erez never planned to work in product. His original path was to become a project manager in construction. He got into product management “by mistake,” landing his first role because he liked the person interviewing him. That mistake turned into a career he now leads at scale, as part of the small team that reports directly to Gong’s Chief Product Officer, Eilon Reshef, alongside colleagues owning AI and foundation, agentic products, and forecasting and deal execution.

Episode Highlights

Fall in Love With the Problem, Not the Technology

Asked what product principle more startups should copy from Gong, Erez didn’t point to a feature or a growth hack:

“Falling in love with the problem and the persona that you’re solving it for, and not with the technologies.”

Startups get built around good technological ideas all the time. The ones that actually make it, in his view, are the ones that go all in on a real user persona and a real problem, and don’t let go until it’s solved.

Everyone Can Build “Their Own Gong” Now. So What’s the Moat?

With agentic AI, Erez is direct about the risk: “everybody can go and take Claude or ChatGPT and create kind of their version of Gong.” So how does an incumbent keep winning?

His answer starts with history, not hype. Gong didn’t bolt AI onto an existing product late in the game, it was built into the platform’s DNA from day one, starting with basic call analysis. On top of that sits what he considers Gong’s real advantage: the ability to ingest and model data properly, at enterprise scale, with the privacy, compliance, and ease of setup that large customers require. That combination, he argues, is much harder to copy than a chatbot wrapper.

Scaling Without Breaking

Erez’s advice to fast-scaling companies is a prioritization problem disguised as a technical one. You’re not building for the fifty-user company, you’re building for the twenty-thousand-user company that hasn’t shown up yet. That means thinking early about how data will perform under load and at scale.

But he’s equally clear about the trap on the other side: overbuilding for scale that doesn’t exist yet can cost you the opportunities in front of you right now. The skill is telling the difference between strategic must-haves and premature tech debt.

Gong Uses Gong: The Secret Weapon

One of the most concrete parts of the conversation is also the most Gong-specific. Product managers at Gong use Gong itself to stay close to customers, searching recorded calls to find every customer asking for a specific feature or using a particular tool, then going to interview them directly.

“You don’t have to wait to hear it from a salesperson or a customer success, you can hear it directly from the customer.”

That includes deliberately listening to lost deals and churned customers, not just wins, to run win-loss analysis and build a sharper picture of what’s actually happening in the market.

Faster Horses vs. Cars

When customers ask for something, Erez’s team treats that as a starting point, not a spec. He leans on the Henry Ford cliché for a reason:

“If I would ask my users what they would want, they would tell me faster horses.”

Gong PMs get plenty of “faster horses” requests: incremental asks to improve one specific feature. Part of the job, as Erez describes it, is explicitly not delivering that. Instead, PMs are expected to have a strong point of view on what’s coming next and build something better than what was asked for. It’s a real product assignment at Gong, not just a philosophy.

Big Dipper: Putting Agent Creation in Users’ Hands

Erez unpacked Gong’s recent Big Dipper announcement, which he frames more as a marketing moment than a single launch event, bundling several features released earlier into one announcement.

The headline is custom agents: the ability for users to create their own agents inside Gong to automate and move work across the platform, and to collaborate with other tools rather than staying siloed inside Gong. The shift he highlights is about control. Previously, Gong curated the agents. Now, agent creation moves into the user’s hands.

The Data Foundation Behind Agentic AI

As every company rushes to support AI agents in customer workflows, Erez’s filter for Gong’s own roadmap starts with a blunt question: is this solving a real pain, or just riding a trend?

His role is making sure Gong’s data infrastructure is ready for agents, not just ingesting more data, but modeling it with the right schema and governance. Skip that step, he warns, and agents don’t have reliable ground truth to work from:

“Without the right schema, the data that you have ingested is not really available to agents, and then they can hallucinate.”

Without governance, agents can also act outside the boundaries of what a user is actually allowed to do. He sees plenty of large enterprise customers eager for agentic AI but still building isolated point solutions internally, not yet ready to deploy it organization-wide, and that gap is exactly the problem Gong is trying to solve for them.

Cheap Iteration, Expensive Evaluation

Erez is candid that predicting which AI trends stick is hard. What he’s confident about is that the cost of iterating has collapsed. A question that used to take a developer days or a week to answer about Gong’s own codebase now takes minutes with AI assistance inside the company.

But he draws a sharp line between iteration and evaluation: testing whether a new AI tool is actually worth adopting is not cheap. Gong runs small “guerrilla teams” of early adopters who test new tools cautiously, with humans reviewing AI output, adjusting the skills and prompts involved, and iterating again, a closed feedback loop he compares to product development cycles from twenty years ago, now applied to AI. If a tool hits a wall and doesn’t deliver real value, the team drops it and moves on.

Nobody Asks a Product Manager to Show Up

Asked about the biggest shift that made him a better product leader, Erez didn’t point to a framework. He pointed to a realization from early in his career, paraphrasing a blog post he read at the time:

“As a product person, nobody asks you to show up and they’ll be fine… if you won’t be there the next day.”

If an engineer doesn’t show up, code doesn’t get written. If sales doesn’t show up, nothing gets sold. A product manager’s absence, by contrast, often goes unnoticed in the short term. For Erez, internalizing that changed how he approached the job entirely: it’s no longer about writing a good spec, it’s about proving, constantly, that you’re solving the right problem, because no one else is going to do that for you.

His advice for PMs who want to grow into senior and director-level roles follows the same thread: expand your mind, and keep asking the strategic questions you didn’t have time to ask when you were heads-down shipping features.

The Bottom Line

You can spin up a prototype of “your own Gong” with off-the-shelf AI in a weekend. What you can’t clone that fast is a data foundation built with the right schema and governance, a product team that treats customer calls as ground truth instead of hearsay, and the discipline to build what customers actually need instead of what they ask for.

Erez’s advice is the practical kind: fall in love with the problem, not the tool, get your data foundation right before you scale agents on top of it, and never let a customer request replace your own point of view.

Watch the full episode to hear how Erez thinks about product strategy inside Gong, why “faster horses” still shapes how his team prioritizes, and what it really takes to keep winning as an incumbent in the agentic AI era.

By Dana Bilder Morozova
Marketing & GTM strategist with a passion for brand positioning, creative formats, and bold product storytelling. At Yuki, I build narratives not around fluff, but around real people – the engineers, data experts, and decision-makers who make the product matter. Drawn to clarity, driven by curiosity.

Table of Contents

Free cost analysis

Take 5 minutes to learn how much money you can save on your Snowflake account.

By clicking Submit you’re confirming that you agree with our Terms and Conditions.

Follow us on LinkedIn

Related posts

Free cost analysis

Take 5 minutes to learn how much money you can save on your Snowflake account.

By clicking Submit you’re confirming that you agree with our Terms and Conditions.

Skip to content