AI raises the stakes for security. Protecting data privacy against other humans or predictable software is one thing; with non-deterministic agents in the mix, the challenge grows exponentially. Anshu Sharma, founder and CEO of Skyflow, joins me to talk about securing sensitive customer data across data stores, models, and agents, as well as some of his biggest lessons as a three-time founder.
Today, raw enterprise data can't simply be fed into a model: it's full of personal information that has to be transformed so it protects privacy while preserving meaning. Skyflow controls what data reaches a model, enforces policy on what an agent is allowed to do with it, and manages data sovereignty across borders.
Anshu has been chasing this problem for over two decades.
At Salesforce, he had to convince early enterprise customers like JPMorgan and Cisco to trust the cloud with their most sensitive data; work that led him to help build Salesforce's identity and app-exchange infrastructure from scratch.
That same trust problem kept resurfacing across two more startups before he founded Skyflow in 2019. “At some point you realize, you know what? There's probably a $100B company to be built just protecting the most sensitive data for customers at all these companies,” he explains.
As a founder, Anshu draws lessons from leaders like Marc Benioff, noting the way companies position themselves against rivals and partners can shape their success in the market. He also argues that enduring companies start with a meaningful problem and leaders who can build a culture around the shared mission to solve it.
We close with his advice for technical founders: learn to respect marketing as a craft. The best sales and marketing comes from understanding your audience so well that it doesn’t feel like sales or marketing at all.
What we covered:
0:00 – Cold open: Anshu on the biggest misconception in AI security
1:00 – What Skyflow does
2:35 – Why AI agents raise the stakes on data privacy
3:55 – Anshu’s story: from Salesforce to Skyflow
8:05 – How AI is reshaping SaaS and infrastructure
10:44 – Skyflow's role in the token economy
15:59 – The biggest misconception about AI security
18:22 – Is cybersecurity a failed industry?
19:08 – How Skyflow's product and go-to-market evolved with the AI shift
20:57 – Building culture in a company competing with OpenAI and Anthropic for talent
22:40 – Lessons from Marc Benioff: beginner's mind and breaking the rules
26:12 – The open-weights debate: trust vs. transparency
30:16 – Advice for technical founders: respect the art of marketing
31:18 – Closing thoughts
Links:
Skyflow: https://www.skyflow.com/