Codos AI

Codos is pitching itself as a virtual Chief AI Officer for mid-sized enterprises, a software layer that runs an AI transformation program end to end instead of leaving it to consultants or internal task forces. The platform interviews employees across an organization, identifies automation opportunities, and deploys agents across functions , with the whole stack designed to run on the customer's own servers rather than in a shared cloud. This launch matters because most enterprise AI spend has yet to translate into measurable P&L movement, and Codos is packaging the diagnostic, prioritization, and deployment work into a single product with an on-prem "company brain" that C-suites can query directly. The team is led by CEO Dima Khanarin, previously a consultant at McKinsey and CEO of the Everclear Foundation, alongside CTO Gleb Sidora, who worked on machine learning at Meta and was a founding engineer at Lightspeed-backed Condukt. The company is backed by a16z speedrun and is targeting midsize companies and enterprise business units of roughly 300 to 3,000 people in fintech, ecommerce, technology, real estate, retail, and services , sectors where headcount-heavy operations create obvious automation surface area. The workflow moves from agentic interviews and data analysis in the first few days, into a pilot automation and initial company brain deployment over the following weeks, then into scaled rollouts across cost centers. For operators evaluating whether to build an internal AI function or buy one, Codos is positioning itself as the packaged alternative, with dedicated forward deployed engineers and on-prem data handling as the main differentiators against a DIY stack.