In a move that underscores the accelerating consolidation in the AI agent space, customer service software giant Zendesk announced this week it’s acquiring Forethought, a pioneering startup that’s been automating customer interactions since 2018. The deal, expected to close by month’s end, represents more than just another corporate acquisition – it’s a bellwether for how established tech companies are scrambling to integrate AI agent capabilities before they’re left behind.
Forethought was years ahead of the current AI agent hype cycle, winning TechCrunch’s Battlefield competition in 2018 – four years before ChatGPT’s debut. The company grew to support over a billion monthly customer interactions for clients like Upwork, Grammarly, and Datadog, raising $115 million from investors including NEA, Sound Ventures, and even actress Gwyneth Paltrow. Zendesk, which went private in a $10.2 billion deal in 2022, says the acquisition will accelerate its product roadmap by more than a year, integrating Forethought’s technology into specialized agents, self-improving AI, and voice automation.
The Infrastructure Race Heats Up
While Zendesk-Forethought represents consolidation at the application layer, the infrastructure supporting AI agents is experiencing its own funding boom. Just days before the acquisition announcement, AgentMail raised $6 million in seed funding to build email services specifically for AI agents. The platform provides API-based email inboxes with features like two-way conversations, parsing, threading, and searching – essentially giving AI agents the same email capabilities humans take for granted.
“We want to give agents the ability to use email in the same way that humans do,” says AgentMail co-founder Haakam Aujla. “What humans use email for is not even communication. It’s your identity. There are several startups trying to build new identity protocols for agents, but our thesis is, let’s just use what already works for humans.”
The company has already attracted tens of thousands of human users and hundreds of thousands of agent users, with user count tripling in one week and quadrupling in February following OpenClaw’s debut. This explosive growth suggests we’re witnessing the early stages of a complete infrastructure stack emerging specifically for AI agents.
Regulatory Headwinds and Ethical Dilemmas
Even as the AI agent ecosystem expands, significant headwinds are emerging. Anthropic, the AI startup behind Claude, faces potential billions in losses after the U.S. Department of Defense labeled it a supply-chain risk. The designation came after Anthropic refused to allow its technology to be used for mass surveillance of Americans or autonomously firing weapons.
The business consequences have been immediate and severe. According to court papers, current and prospective customers are demanding new contract terms or backing out of negotiations entirely. This regulatory friction highlights a growing tension between AI companies pursuing commercial applications and government entities with different priorities.
In a remarkable show of industry solidarity, more than 30 employees from OpenAI and Google – including Google DeepMind chief scientist Jeff Dean – filed an amicus brief supporting Anthropic’s legal fight. “The government’s designation of Anthropic as a supply chain risk was an improper and arbitrary use of power that has serious ramifications for our industry,” Dean stated in the brief.
The Creative Frontier: Promise and Peril
Beyond customer service and infrastructure, AI agents are pushing into creative domains with mixed results. The debut of AI-generated “actor” Tilly Norwood and her song “Take the Lead” has sparked controversy about authenticity and artistic value. The production involved 18 contributors, yet the song’s theme – an AI character’s struggle against human skepticism – highlights the fundamental disconnect between AI-generated content and human experience.
This tension isn’t limited to entertainment. SAG-AFTRA, the actors’ union, issued a statement criticizing such creations: “‘Tilly Norwood’ is not an actor; it’s a character generated by a computer program that was trained on the work of countless professional performers – without permission or compensation. It has no life experience to draw from, no emotion and, from what we’ve seen, audiences aren’t interested in watching computer-generated content untethered from the human experience.”
What This Means for Businesses
For enterprises considering AI agent adoption, several key takeaways emerge. First, the infrastructure is maturing rapidly, with specialized services like AgentMail filling gaps in the ecosystem. Second, regulatory uncertainty remains a significant risk factor, particularly for companies working with government entities or in sensitive domains. Third, while AI agents can handle routine tasks efficiently, creative applications face skepticism about authenticity and value.
The Zendesk-Forethought deal suggests that for established companies, acquiring AI agent capabilities may be faster than building them internally. But as the Anthropic case demonstrates, navigating the regulatory landscape requires careful consideration of ethical boundaries and potential business impacts.
As AI agents move from novelty to necessity, the industry stands at a crossroads between rapid commercialization and responsible development. The coming months will reveal whether these competing priorities can be reconciled – or whether they’ll pull the ecosystem in fundamentally different directions.