Integrating Gen AI into Product and Organization — with Ohad Peri (Bllink)
Host: Eyal David · Guest: Ohad Peri · About the host
According to Ohad Peri, CTO and a founding-team member at Bllink, Gen AI doesn't make people redundant — it makes redundant those who don't use it; a developer or designer who adopts it can do three jobs in the time of one. Bllink built it into every product layer, improving Time to Merge by up to 50%.
In this episode
- Bllink's first feature (onboarding a building through an AI chat instead of a human rep) failed — not technically, but psychologically: building-committee managers who entrust money want to see there's a human behind the process. But the failure produced the AI infrastructure that today powers the automation of customer support.
- The AI infrastructure was built modular and split into three: a chat layer (third-party), a set of actions that can be run in the system, and the AI layer that decodes the user's intent and runs the appropriate function with their permissions — so the GenAI model can be swapped out without losing logic or data.
- Coding with AI is the most significant case: tools like GitHub Copilot, Cursor, and Vercel's V0 produce up to a 50% improvement in Time to Merge, and features ship at a faster pace that translates directly into market value.
- The guiding line: the technology "doesn't make people redundant, it makes redundant the people who don't use this technology" — a designer who uses Gen AI can do three jobs in the time of one.
- The optimistic vs. dystopian vision: instead of mass unemployment, AI may dramatically increase the number of problems the world solves — solo entrepreneurs will solve problems that today require giant companies, and giant companies will tackle problems no one touches today.
- In a small startup, "measuring efficiency is terribly inefficient" — Bllink didn't measure the exact savings from their AI tools, but they see the efficiency in practice and choose to run forward.
Listen to this episode
Episode Description
Ohad Peri, CTO and a founding-team member at Bllink — a startup digitizing building-committee (va'ad bayit) payments in Israel — joins for a conversation entirely about integrating Gen AI into a product and an organization. The episode moves from the failure of the first feature (a chat-based onboarding that failed, of all things, for a psychological reason) to the modular infrastructure that was born out of it, and from there to a broad survey of case studies from the market: accelerated code development, design, the automotive industry, and the big question — is the enormous investment in AI economically justified?
Full transcript with speakers and timestamps
Selected quotes
“In a small startup there's a saying I like to use, measuring efficiency is terribly inefficient.”
“It doesn't make people redundant, it makes redundant the people who don't use this technology.”
“You have to say no, and when you say no, you need to know why you're saying no, you need to explain it well, and you need to explain it nicely too.”
Questions & answers
What does Bllink do?
Bllink is a startup that solves the problem of building-committee (va'ad bayit) payments in Israel — a huge market where, until a few years ago, about 90% of payments were made by check. Bllink lets tenants pay through a simple web interface, including recurring payments, and spares the management companies the logistical and operational pain.
Why did the first AI-based feature fail?
The idea was that a building committee could onboard a new building through an AI chat with no human rep. The assumption turned out to be wrong: in a complex onboarding that touches money, people want a human to reassure them. But building it brought into Bllink the AI infrastructure they use today.
What is the Sequoia research mentioned in the episode?
Sequoia, one of the world's largest VC funds, published an article titled "The $600 Billion Question" — the amount invested in developing AI (mainly Foundation Models) — asking whether the enormous investment is economically justified. Ohad answered by comparing it to the build-out of the U.S. electrical grid infrastructure 60 years ago, where costs ran to about 10% of GDP.
Which AI tools does Bllink actually use?
GitHub Copilot (for about a year now) for code autocomplete, Vercel's V0 for building front-ends from text, Cursor as an AI-based IDE, and DALL·E to generate creatives and assets for ads. They work in Jira for task management and mention Notion as an example of AI that works well.
What are the most important PM skills, in Ohad's view?
First — transparency and communication: breaking down the high-level logic for developers so they understand why they're doing what they're doing. Second — the ability to say "no": resources are always limited, and you need to know why you're saying no, explain it well, and do it nicely too.
Related Episodes
How to Build AI at the Edge of the Device — Ori Katz, Hailo
00:40:49 · March 5, 2026
Ori Katz, who leads the AI SoC domain at Hailo, tells the story of how an Israeli chip company became a serious player in AI at the edge — smart camer...
Tom Mayblum | MIND | When Is a Customer Conversation Worth More Than Data?
00:33:39 · February 9, 2026
Tom Mayblum, VP Product at MIND, in a short tachles episode on when a direct customer conversation is worth more than looking at the data — and when y...
Keren Wexler | Optimove | Product-Team Processes That Create Big Impact
00:44:13 · December 23, 2025
Keren Wexler, Senior Director of Product at Optimove, on product, teams and decision-making — and the point where process stops being bureaucracy and...
Why "I Didn't Hear Anything New" Is the Product Leader's Real Win
00:04:07 · December 21, 2025
Noa Ganot breaks down why product managers fail to move the executive team: it's not a shortage of slide decks, it's the lack of a story that explains...