Sharing real-world enterprise AI use cases, straight from conversations with early adopters actually deploying AI and enterprise agents. No fluff. Just what’s working, what’s not, and what’s next in AI & Agents for business. 🍺
This week, my colleague Stephen Stouffer and I hosted a LinkedIn Live session to unpack the real-world state of AI adoption & share some production Agents. (Recording below).
Over the past few months, Stephen and I have been on the road—London, San Francisco, New York, Detroit, and most recently, Vegas for the Gartner Application Innovation Summit. In every city, we’ve been meeting with enterprise leaders, technologists, and teams on the front lines of automation and transformation initiatives.
And what we’re hearing might surprise you.
The narrative around AI in the media tends to swing between extremes, either AI is poised to replace entire workforces tomorrow, or it’s overhyped vaporware. But the actual conversations we’re having paint a more nuanced, and frankly more encouraging, picture.
Here were our main insights from conversations in the field:
1. General AI curiosity, but low hands-on experience
AI remains new territory for everyone—most attendees at the Gartner App Summit and similar events are in early stages of understanding and adopting AI.
Despite high interest, very few people have actually built agents—out of ~300 attendees, only ~15 raised their hands when asked if they had.
2. Anxiety and pressure to “Catch Up”
A common theme was anxiety: many feel behind in their AI journey due to media, leadership, and market pressure.
Leadership (CEOs, boards) often mandates an “AI strategy,” but frontline teams are unclear where to begin.
3. Overreliance on “Easy Button” tools
Tools like Microsoft Copilot, Google Gemini, and Glean are being switched on as quick wins.
These are often treated as AI adoption milestones, but in reality they’re table stakes
4. Fragmentation and siloed solutions
Teams are deploying AI features within individual SaaS tools, leading to fragmented, narrow use cases.
This recreates the early SaaS sprawl—now happening with AI (Agent Sprawl).
5. From Chat to Action
The real next step is moving beyond chat-based Q&A to autonomous execution—agents that take action inside enterprise systems.
Tray is investing in this “Phase 2” of AI: not just answering, but doing.
6. Security, Compliance, and Regulation are Major Blockers
Many companies hesitate due to lack of compliant LLMs (e.g., HIPAA, SOC2, GDPR).
Regional data hosting, especially for EU-based companies, is a hard requirement and not widely supported by AI tools today.
7. Protocols like MCP and A2A are early
New standards like A2A (agent-to-agent) and MCP are being discussed, but practical use cases are still few and far between.
8. People are fed up with Salesforce
“Salesforce is the new Oracle” their words not mine. : )
People are growing tired of the “necessary evil” SFDC machine. The once all powerful seems to be grasping to their market share and making moves that negatively impact their Customers.
A prime example of this was their recent move to block other apps from searching or storing Slack messages.
For your viewing pleasure:
I started this newsletter because I am frustrated by the lack of tangible Enterprise AI use cases in the market.
I have the opportunity to speak to hundreds of tech and systems leaders and fundamentally believe that AI and Agents will change the way businesses operate. My goal is to help share how.
-Nate G
PS - I went to Regionals in “Power of the Pen” in 8th grade so I consider myself a fairly prolific writer.