"I Didn't Go In, I Went Up": How to Turn a Service into a Product | Orly Amrany (Justt) - Product Builder Podcast episode with Eyal David
Updated:

"I Didn't Go In, I Went Up": How to Turn a Service into a Product | Orly Amrany (Justt)

Host: · Guest: Orly Amrany · About the host

Orly Amrany, CPO at Justt (a startup that helps merchants fight chargebacks), turned a manual, template-based operation into a product that generates defense documents on its own. In her view, the way to turn a service company into a product company is not to dive into the thousand edge cases the team carries in their heads, but to go up a level: define a vision and a machine, and ask the team to bring the cases that contradict it. The move took eight months, included migrating 100% of customers with A/B testing and a way back, and required a CEO who went all in.

In this episode

  • To turn a service into a product, you don't dive into the thousand edge cases the team carries in their heads. You go up a level, define a vision and a machine with basic logic, and ask the team to bring real cases that contradict it.
  • The difference between automation and product: automation gives people better tools to do the same work, while a product makes the manual work unnecessary. Justt doubled its ARR every year while manual work kept shrinking.
  • The hard part isn't the technology but the DNA of a service organization. It took a CEO who goes all in, buy-in from the domain experts, and giving the operations people veto power.
  • Responsible migration: a testing plan before the product is ready, a 50-50 A/B test on a large customer, a way back to the old templates, and the first goal is not to break anything. After that you can only improve.
  • In the end everything comes down to data: what most affects the chance of winning a case is which information you bring, and AI helps especially with extracting unstructured evidence such as photos and support conversations.
  • In the AI era precision is back: when the one coding is a machine, you again need to write a precise definition with all the context, but prioritization and driving things through the organization still stay with people.

Episode Description

Orly Amrany, CPO at Justt, came to product from development (the IDF's Mamram unit and the Technion) and has more than 15 years in product management, including running monday's rebranding project from dapulse. She explains what chargebacks are and why merchants lose up to one percent of their turnover to them, and describes how, when she joined Justt, she found an operation that manually adapted templates for every customer, error-prone and impossible to scale. Instead of giving the operation better tools, she built a machine: she defined basic logic, made the team bring cases that contradicted it, and got buy-in from the CEO and the co-founder. After eight months and a careful migration with A/B testing and a 120-day wait for results, Justt moved to a "dynamic document" instead of templates, a differentiator that became its elevator pitch. Orly also talks about where AI really helps today (documents, unstructured evidence, onboarding), why everything comes down to data, why precise definitions are back in the AI era, and how Justt is already monitoring purchases made by AI agents ahead of agentic commerce.

Chapters

  1. 0:00 · Opening: monday's rebranding project and the secret of execution
  2. 2:53 · Will we see more company spin-offs in the AI era?
  3. 5:20 · Who is Orly: from development and Mamram to product management
  4. 6:30 · How the product role changed, and why precision is back now
  5. 8:18 · What product managers talk about today: where AI helps and where it doesn't
  6. 11:00 · What Justt is: the war on chargebacks
  7. 13:40 · Three and a half years in: scaling without inflating operations
  8. 18:18 · The AI layers: document generation, evidence quality and unstructured data
  9. 20:30 · The bottleneck: a service that doesn't scale
  10. 22:00 · The aha moment: going above the complexity instead of diving in
  11. 23:34 · Eight months to stand up a new machine
  12. 24:01 · The real challenge: working with people who hold all the knowledge in their heads
  13. 26:36 · The CEO's buy-in and why they went all in
  14. 28:00 · Migration plan, A/B testing and a 120-day wait
  15. 29:14 · The results, and why once you haven't broken anything you can only improve
  16. 32:22 · The "dynamic document": the differentiator they branded
  17. 33:50 · AI-first: automatic onboarding and why it all comes down to data
  18. 38:01 · How to stay relevant, for product and for the organization
  19. 43:14 · When AI and when logic: visibility into everything the system does
  20. 45:40 · What's next: a roadmap for recovery and agentic commerce
  21. 47:31 · Wrap-up: from service to product
Read the full transcript
Full transcript with speakers and timestamps

Selected quotes

“Why do I need to go inside, when I can not go in, go up a level and say, fine, whatever logic you have in your heads, I'll sort it out.”
— Orly Amrany
“Bring use cases that contradict it, that's the goal.”
— Orly Amrany
“I said, first of all, don't break anything. Now, after we managed to do that, there's an opportunity to improve a lot more.”
— Orly Amrany

Questions & answers

What does Justt do?

Justt helps merchants, anyone who sells online, fight chargebacks: it submits a defense document with evidence, such as a shipment tracking number, to the issuer and recovers the merchant's money. The company was founded in 2020, has raised 100 million dollars in three rounds, has about 120 employees and works with companies such as Meta, Spotify and DoorDash.

What is a chargeback and why does it hurt merchants?

A chargeback is when a customer asks their card issuer for a refund, and the default is to take the money from the merchant until proven otherwise. It can reach half a percent to one percent of turnover, plus a fee of about 15 dollars per case, and too high a rate can lead Visa and Mastercard to shut the merchant down.

How did Orly Amrany turn Justt's operation into a product?

She sat with the operations team, realized that the underlying logic was rule-based, and built a machine instead of continuing to adapt templates by hand. She asked the team to bring cases that contradicted the vision, solved them together with them, and within eight months moved 100% of customers to the new system.

What is a Dynamic Document?

Instead of fixed templates, Justt's system builds each case its own defense document based on the customer's specific claim and the available evidence. It's the differentiator Justt branded and turned into its elevator pitch, and competitors later copied it.

How will agentic commerce affect chargebacks?

When AI agents buy on behalf of users, there will also be chargebacks on those purchases, and it's still unclear who is responsible: the agent, whoever configured it, or the user. Justt is already monitoring traffic to identify agent purchases and collecting the data, so it can adapt quickly when regulation arrives.

People & resources mentioned

People

Companies & orgs

Tools

  • Gong
  • n8n
  • Claude
Chargebacks From service to product Productization Migration Dynamic Document Agentic Commerce Fintech Justt