Flight Path: How AI Startups Can Partner With Enterprises and Win

Flight Path: what enterprise leaders actually want from AI startups

Enterprise AI strategy has shifted. Buyers no longer want another vendor pitching features — they want partners who understand their problems, integrate into existing systems, and earn trust before scaling. Pilot programs that once stalled for quarters now move to production in weeks, and the startups that win these partnerships are the ones who show up with domain expertise, not just demos. (For a data-driven look at how enterprise AI adoption is accelerating, see Wing's The State of AI in the Enterprise.)

Flight Path is Wing's video series built around that reality. In each episode, we sit down with a senior technology leader at a major enterprise — from Salesforce to Waymo to Politico — and ask a simple question: what do you actually look for in an AI startup partner? The answers are frank, specific, and immediately useful for founders trying to land and expand their first enterprise deals. I think the patterns that emerge across all six conversations are consistent enough to constitute a framework — and specific enough to act on.

Contents

• What enterprises look for in AI partnerships
• Salesforce's Brad Arkin on enterprise scale and startup partnerships
• StubHub's Bob Kupbens on how AI can uplevel live experiences
• Waymo's Vincent Vanhoucke on embedding AI into physical systems
• What Paramount wants from AI startups
• AI should assist, not replace journalists: Politico's approach to AI
• The University of Chicago's chief information officer on why universities make ideal AI test beds
• Key takeaways for founders

What enterprises look for in AI partnerships

Six conversations with enterprise leaders surface a remarkably consistent set of expectations. These are not abstract wishlists — they are the decision criteria that determine whether a startup gets a second meeting or a polite pass.

• Domain understanding over generic capability. Paramount's Phil Wiser measures AI success through organic adoption, not flashy demos. Politico's Beth Diaz insists AI tools must respect the nuances of journalism. Enterprises reward startups that learn the buyer's world before building.
• Security and trust as prerequisites. Salesforce's Brad Arkin describes how startups must raise security assurance without slowing teams down. The University of Chicago's Kevin Boyd flags compliance-driven buying as a defining constraint in regulated institutions. Trust is not a feature — it is the foundation.
• Clear problem-solution fit. StubHub's Bob Kupbens warns against inventing new problems. Waymo's Vincent Vanhoucke emphasizes that AI in physical systems demands rigorous validation and simulation. Enterprises adopt AI that solves problems they already have, not problems a startup wishes they had.
• A compelling origin story. Arkin highlights the importance of a genuine "why this, why now" narrative. Buyers want to know why a founding team is uniquely positioned to solve this problem — and why the timing is right.

Salesforce's Brad Arkin on enterprise scale and startup partnerships

This episode of Flight Path features Brad Arkin, Chief Trust Officer at Salesforce, who walks through what enterprise-grade startups must get right if they want to catch the attention of a Fortune 500 company. He dives into how automation and AI increase the velocity of problem solving, how to raise security assurance without slowing teams down, and the importance of a genuine “"why this, why now”" origin story.

• Watch the video
• Read the Q&A

StubHub's Bob Kupbens on how AI can uplevel live experiences

In this installment of Flight Path, Bob Kupbens, CEO of StubHub International, shares how live-event platforms are using AI not to invent new problems, but to solve the ones that already matter. He explains why brands must deeply understand what fans love and expect, why replacing humans with bots can backfire, and where real-world automation elevates the experience.

• Watch the video
• Read the Q&A

Waymo's Vincent Vanhoucke on embedding AI into physical systems

Vincent Vanhoucke, VP of Robotics at Waymo and former head of Google Brain, explains what happens when AI leaves the screen and enters the physical world. He discusses the organizational culture required to ship AI safely at scale, and why autonomous driving treats safety as a design constraint rather than an afterthought. He also names what he calls the "AI and..." challenge — the hard work of validation, simulation, and continuous monitoring that separates research prototypes from production systems.

• Read the Q&A

What Paramount wants from AI startups

Phil Wiser, Chief Technology Officer at Paramount, offers a candid look at how a major media company evaluates AI partners. He argues that real AI success shows up as organic adoption across teams — not in polished demos. Wiser explains why Paramount builds franchise-aware creative AI tools, how the company avoids scattershot use cases by focusing on high-impact problems first, and why a partnership-first posture matters more than a buy-versus-build debate.

• Read the Q&A

AI should assist, not replace journalists: Politico's approach to AI

Beth Diaz, Senior Vice President of Revenue Operations and Strategy at Politico, lays out a clear boundary for AI in newsrooms: assist, never replace. She walks through how Politico evaluates AI vendors with a sharp eye on security and privacy, uses AI for audience analysis to sharpen editorial focus, and recognizes the hard limits of automation in investigative journalism. For startups selling into media, her perspective is a roadmap for earning trust in a skeptical industry.

• Read the Q&A

The University of Chicago's chief information officer on why universities make ideal AI test beds

Kevin Boyd, Chief Information Officer at the University of Chicago, makes the case that universities are among the best proving grounds for AI startups — if founders know how to navigate them. He explains how institutions evaluate AI tools through a compliance-driven lens, why multi-stakeholder governance shapes every buying decision, and what maturity of solution actually means to a chief information officer managing thousands of users across research, administration, and student services.

• Read the Q&A

Key takeaways for founders

Six conversations with enterprise leaders converge on five principles founders can act on immediately:

1. Lead with the problem, not the technology. Every enterprise leader in this series values startups that deeply understand the buyer's domain and solve an existing pain point.
2. Treat security as a foundation. Whether you are selling to Salesforce or the University of Chicago, security assurance and compliance readiness are table stakes — not differentiators.
3. Measure success through adoption, not demos. Paramount's Wiser is explicit: organic adoption across teams is the real signal. Build for daily use, not quarterly business reviews.
4. Move quickly from pilot to production. Enterprise AI adoption accelerates when startups prove they can scale within existing architecture. A narrow, high-impact first use case earns the right to expand.
5. Tell a genuine origin story. Buyers want to know why your team, why this problem, and why now. A compelling narrative builds conviction faster than a feature list.