10 million customers in Kavak's database
Kavak has 10 million customers in its database, each with AI agents assigned to maximize their lifetime value.
Snapshot · The a16z Show
The Self-Improving Company | Kavak's AI Playbook
Explore episode
Aug 10, 2026
Where this was said
The Architecture: One Agent Per Customer
At 7:15 · chapter starts 4:00
Angela asks Alejandro to ground the listener in what actually happens when a customer arrives. The answer is striking: a dedicated agent is spawned specifically for that customer, running in its own virtual machine. It remembers years of interaction history — a webpage visit, a call two years ago — formulates a long-term strategy, and relentlessly works to maximise lifetime value across all of Kavak's products over time. Alejandro then explains the three foundational decisions that made this possible. First, Kavak had to resist the instinct to simply hand ChatGPT to employees and instead rebuild its entire API and system layer so agents could actually act. Second, it made the bold bet that agents could become superhuman — not just helpful, but genuinely better than the best human ever hired. Third, it reoriented the company's success metrics from transactional (cars sold, parts ordered) to relational, assigning agents to all 10 million customers with a mandate to maximise lifetime value. [1] — Alejandro Maza Ayala "Most companies build task-specific agents. Kavak builds one agent per customer, with its own virtual machine, long-term memory, and a singl…" 04:00
Agent Per Customer, Not Per Task
Most companies build task-specific agents. Kavak builds one agent per customer, with its own virtual machine, long-term memory, and a single goal: maximize that customer's lifetime value over years. The agent wakes up, works, sets an alarm, and comes back — it never forgets.
Why Kavak Blew Up Its Org to Rebuild Around Agents
Simply giving employees ChatGPT changes nothing. Kavak's first move was to rebuild its APIs, systems, and feedback loops so agents could actually do the work. The second bet: agents could be superhuman — outperforming the best human hire on every metric that mattered.
96% of Interactions, 95% of Transactions: Kavak's Agent Scale
96% of all customer interactions and 95% of all transactions at Kavak are now handled by AI agents. Between 100,000 and 200,000 agent instances wake up every day, each with its own virtual machine, working anywhere from 3 minutes to 3 days before setting an alarm for their next task.
Similar snapshots
$13K MRR with $0 marketing spend
Bhanu grew SiteGPT to $13,000 monthly recurring revenue entirely through organic channels, spending nothing on paid marketing.
1 million website visitors total
More than 1 million people have visited SiteGPT's website since launch in March 2023, all through organic channels.
90% of Google traffic from free tools
Approximately 90% of SiteGPT's Google search traffic comes from the free tools Bhanu built, not the main product pages.
Sold first SaaS for $250,000
Bhanu sold his first SaaS product, Feather, for $250,000 so he could focus fully on the faster-growing SiteGPT.
$500K total revenue since launch
SiteGPT has generated approximately $500,000 in total revenue since its launch in March 2023.
SiteGPT LTV ~$1,700–$1,800
The average customer lifetime value for SiteGPT is approximately $1,700 to $1,800, which Bhanu considers unusually high.
50,000 monthly visitors
SiteGPT receives around 50,000 visitors per month, of which about 200 convert to leads and 60 start free trials.
SiteGPT reached $10K MRR in month 1
SiteGPT hit $10,000 MRR within its very first month of launch, driven largely by early traction in the AI chatbot space.
Stuck at $2K MRR despite downloads
Despite strong download numbers, PropGPT could not push past $1,000–$2,000 MRR due to poor product retention.
