Scaling Retail Shelf Recognition with IBM
Key Points
- Maksim Morozov, CEO of an Eastern‑European intelligence‑retail tech firm with operations in Finland and Russia, highlights the persistent “out‑of‑shelf” problem in brick‑and‑mortar stores.
- Missing items, incorrect pricing and outdated promotions cost the retail industry over $500 billion each year, prompting the company to develop a visual‑recognition platform that can instantly flag stock‑outs.
- The solution must handle massive image‑processing loads—from tens of thousands to tens of millions of photos monthly—and be easily scalable across new regions and markets.
- To meet these reliability and big‑data demands, the firm selected IBM’s infrastructure and analytics services as the backbone for its global rollout.
Full Transcript
# Scaling Retail Shelf Recognition with IBM **Source:** [https://www.youtube.com/watch?v=2hlvVAOdxw4](https://www.youtube.com/watch?v=2hlvVAOdxw4) **Duration:** 00:01:36 ## Summary - Maksim Morozov, CEO of an Eastern‑European intelligence‑retail tech firm with operations in Finland and Russia, highlights the persistent “out‑of‑shelf” problem in brick‑and‑mortar stores. - Missing items, incorrect pricing and outdated promotions cost the retail industry over $500 billion each year, prompting the company to develop a visual‑recognition platform that can instantly flag stock‑outs. - The solution must handle massive image‑processing loads—from tens of thousands to tens of millions of photos monthly—and be easily scalable across new regions and markets. - To meet these reliability and big‑data demands, the firm selected IBM’s infrastructure and analytics services as the backbone for its global rollout. ## Sections - [00:00:00](https://www.youtube.com/watch?v=2hlvVAOdxw4&t=0s) **Scaling AI Shelf Recognition** - CEO Maksim Morozov describes how his Eastern‑European retail‑technology company tackles costly out‑of‑stock issues by processing massive volumes of shelf‑photo data for real‑time product recognition, and why they selected IBM’s robust big‑data infrastructure to scale the solution globally. ## Full Transcript
my name is Maksim Morozov I'm chief
executive of intelligence retail company
so we're a technology company from
Eastern Europe we have a presence in
Finland and Russia there's a big
challenge still in in brick-and-mortar
retail we are all shoppers right and
when we visit stores we expect a
well-organized shelf so it means that
all the products should be at the shelf
so prices should be correct so
promotions are not outdated and that's a
big problem number one problem for
retail now it costs more than 500
billion every year if the product is
missing for example if the product is
out of shelf and you don't know it every
minute you lose money that's what we try
to fix it means that we need more
recognition power on our servers for
example if the customer starts in one
region and they start to send us like
100,000 follows per month they can
easily roll out the solution to to other
countries to other regions and increase
in this number up to some Millions we
now know that we can support easily any
projects any size of the customer and we
can easily scale rollout to other
countries and other regions we are not
afraid now if the customer send us you
know ten millions photos to be
recognized we know that infrastructure
is allows us to do and end to scale our
recognition power we need a reliable
infrastructure and big data solutions
that's why we at the end we chose IBM