Key points
- Surojit Chatterjee co-founded Google's Mobile Ads team, grew it to $50bn, and now leads Ema, a universal AI employee for enterprises.
- Ema stayed in stealth for a year, building with customers as design partners before announcing its fundraise.
- Chatterjee says success creates norms that block innovation, and culture degenerates unless leaders are intentional and willing to be unpopular.
How conviction built Google's $50bn mobile ads business
Surojit Chatterjee's career is a case for conviction over consensus. He grew Google's mobile ads from zero to $50bn by ignoring the company's own statistical-significance rule, and watched the culture turn inward as the company scaled. Ema, his enterprise AI company, is the same bet applied to a new market, an AI employee that absorbs the repetitive work inside large companies.
Takeaways
- Ema is building a universal AI employee. It works alongside human staff on repetitive enterprise tasks. The product was built with several customers as design partners, and the company stayed in stealth for a year without announcing its fundraise or incorporation.
- Enterprise talent wastes time on repetitive work. Running large teams at Google, Flipkart and Coinbase, Chatterjee was bothered that a large share of talented people's time went to repetitive tasks. Ema is meant to automate that work so people can spend their time on creative work instead.
- Conviction beat Google's launch rules. Google required statistical significance before any launch, which with few mobile users meant nine months to a year per feature. Chatterjee turned click-to-call ads on for 50 percent of users and offered to take the blame. Mobile ads reached $50bn in revenue.
- Success builds the norms that block innovation. Google's rules assumed the user numbers that success had brought. Chatterjee says smart people often argued his ideas were impossible because someone had tried and failed before, and he treats vehement opposition as fuel rather than a stop signal.
- Culture decays unless leaders are deliberate. At Google, free expression went to an extreme and people spent their days on things outside the mission. Chatterjee says running a company is not a popularity contest, leaders cannot please everyone, and they have to take stances that make them unpopular.
- India was mobile-first before the West. At Flipkart, 90 percent of e-commerce GMV came from phones against 80 to 90 percent desktop in developed markets. Serving a billion people with few cards and poor roads meant designing thin trucks that could reach villages through small lanes and alleys.
- Founders should build and ignore noise. Chatterjee says Twitter can absorb a whole day with other people's wisdom, and conviction has to be your own. His advice to AI founders is to focus on building, because a great product makes everything else fall into place.
What Ema predicts next
- Every enterprise will run on a conversational operating system, with human teams and AI employees working side by side rather than AI used as a tool.