Augment has acquired Merlin, a stealth AI company focused on wholesale distribution. The move gives the supply chain AI platform its first foothold in the more than $8 trillion U.S. wholesale distribution sector. It’s an industry that still relies heavily on legacy systems and spreadsheets.
The deal brings distribution industry veteran Alex Moazed to Augment as president of wholesale distribution, alongside co-founders Nick Johnson and John Schumacher, former head of AI at Grainger.
A reason behind the acquisition? Customer demand. Freight brokers and shippers kept asking Augment CEO Harish Abbott if the platform worked with distributors, his biggest customers.
“When we looked at what Alex had created, two things immediately stood out: expertise in a specific distribution domain that would have taken us years to replicate, and a data philosophy that we completely share: Customer data remains isolated, period,” Abbott said. “He never trains models for anyone else.”
Moazed co-founded Merlin after leading Applico Capital, the first venture capital fund dedicated to the distribution industry, where he evaluated more than 700 supply chain technology companies. Applico participated in Augment’s seed round in January 2025.
“When the acquisition talks got serious, the cultural and technological fit was already obvious,” Moazed said.
The case for purpose-built delivery AI
B2B distribution is extremely decentralized, even more so than freight transportation. Enterprise distributors lose margin due to operational complexity driven by fragmented systems. This fragmentation creates manual processes and inconsistent product language across multiple enterprise resource planning (ERP) systems.
“Sales reps spend 30 minutes entering a purchase order that a customer has already written,” Moazed said. “Counter reps come in mid-quote, the ERP crashes, and they start over. Branch managers call suppliers every two weeks just to check on open orders.”
A single company can run six different ERPs. A quote can take two minutes per line on thousands of lines daily. Generic AI tools fail to handle unit conversions, customer-specific SKU language, and workflows with many exceptions. Many also train shared models with proprietary dealer data without returning value to customers.
“Reseller data is a competitive moat – their resistance to shared data AI models is not an objection to overcome, it is a legitimate business concern,” Moazed said.
Augment’s platform, Augie, uses agent AI to follow standard operating procedures written in natural language instead of custom code.