The New Rules of Enterprise Software with Steven Sinofsky

Watch on YouTube ↗  |  July 07, 2026 at 14:50  |  1:01:09  |  a16z
Speakers
Seema Amble — Partner, a16z Enterprise Team
Steven Sinofsky — Host

Summary

The podcast discusses how AI agents and headless software are changing enterprise software access, with APIs and agentic workflows reducing reliance on human UIs. The speakers argue that systems of record like SAP, Salesforce, and Workday remain sticky because business logic, exceptions, compliance, and workflows are deeply embedded. They see opportunities for AI software startups to build action and context layers on top of incumbents, while warning that incumbents may struggle to build great agents themselves. Startups should target gaps between incumbents and bridge organizational functions.

  • Headless software means agents access data via APIs/MCP rather than a human UI.
  • Sticky enterprise software is driven by workflows, muscle memory, compliance, and single sources of truth.
  • SAP and similar systems of record are hard to replace because they codify business rules and customizations.
  • Exception handling is the core challenge for enterprise AI agents.
  • Incumbents are likely to bolt AI onto existing products rather than disrupt themselves.
  • AI software startups can win by adding action, context, and data-capture layers.
  • Vertical software for construction, manufacturing, and physical-world data is an opportunity.
  • Internal network effects and tools that bridge organizational functions are promising categories.
Ideas
Seema Amble Partner, a16z Enterprise Team 10:03
Enterprise software stickiness remains durable.
Enterprise software is durable because it is built around human workflows, frequent read/write access, downstream processes, undocumented SOPs, muscle memory, cross-department dependencies, and the need for a single source of truth. This stickiness persists even as agents change how data is accessed.
IGV HIGH
Seema Amble Partner, a16z Enterprise Team 10:34
Salesforce remains sticky system of record.
Salesforce is sticky because sales reps and managers have muscle memory, VPs mandate it, finance and marketing rely on its data, and it serves as a single source of truth and enforcement mechanism for go-to-market. Even with headless/agent access, the system of record remains valuable.
CRM MED
Steven Sinofsky Host 13:41
Outlook stickiness protects Microsoft enterprise moat.
Microsoft Outlook is deeply sticky because of delegate access, multi-owner calendars, and recurring meeting exception handling; enterprises like GM will not displace hundreds of thousands of seats over such embedded workflows, reinforcing Microsoft's enterprise moat.
MSFT MED
Steven Sinofsky Host 15:41
SAP is sticky and not replaceable.
SAP is the ultimate sticky software because it codifies a company's business rules, regulatory requirements, and customizations; replacing it with a Postgres database and APIs is unrealistic, and removing SAP from large enterprises would destroy how they operate.
SAP HIGH
Seema Amble Partner, a16z Enterprise Team 49:22
Workday API limits reinforce switching costs.
Workday has APIs but limits data extraction and documentation, making it very difficult for customers to cleanly extract all data or operate without Workday; this closedness reinforces switching costs and protects it from disintermediation.
WDAY MED
Seema Amble Partner, a16z Enterprise Team 51:17
AI software is a major investment area.
AI software is an attractive investment area because startups can build alongside SAP/Salesforce, add visibility and action layers, run agents on existing data without replacing backend logic, and create new systems of record by collecting voice, recordings, transcriptions, and documents. This is why a16z continues investing in AI software.
AI software HIGH
Seema Amble Partner, a16z Enterprise Team 52:40
Agentic CRM/sales software has major upside.
Startups can win in CRM/sales by moving beyond data collection to an agentic action layer: prioritizing leads, flagging churn risk, sending outbound, observing responses, and collecting benchmark data on what works by region/context. This creates a new data exhaust and system of engagement.
AI-SECTOR MED
Seema Amble Partner, a16z Enterprise Team 53:30
Vertical physical-world AI software is attractive.
Vertical software for the physical world, such as construction and manufacturing, is attractive because field data from humans and machines has historically been hard to capture; AI agents can now observe and pull that data back into software, creating new systems of record and optimization loops.
Manufacturing software Construction software MED
Steven Sinofsky Host 59:45
AI bridging org functions is opportunity.
Software that leverages AI to bridge functions within an organization—such as IT and finance for cloud budgeting—can create internal network effects and a new enterprise software category, which is a huge opportunity.
AI-SECTOR MED
Up Next
2:24:17
Bloomberg Surveillance 7/7/2026
Bloomberg Markets · Jul 07
27:23
If You Believe in AI, You Have to Bet on This
Bankless · Jul 07
3:03
Eurasia Group president on what to expect at NATO summit
CNBC · Jul 07
6:55
Finland's Stubb Says Ukraine Needs More Patriot Missiles
Bloomberg Markets · Jul 07
3:09
Market Open: Dow Adds to Record; Oil Rises on Hormuz Strikes • 7/7/26
CNBC · Jul 07

This a16z video, published July 07, 2026, features Seema Amble, Steven Sinofsky discussing IGV, CRM, MSFT, SAP, WDAY, AI software, AI-SECTOR, Manufacturing software, Construction software. 9 trade ideas extracted by AI with direction and confidence scoring.

Speakers: Seema Amble, Steven Sinofsky  · Tickers: IGV, CRM, MSFT, SAP, WDAY, AI software, AI-SECTOR, Manufacturing software, Construction software