Forethought

Most customer work still happens in browser-based systems—internal dashboards, vendor portals, and tools without APIs. When something needs to be updated, someone still has to sign in and do it manually. With Browser Agent this all changes. Browser Agent uses a system the same way a support rep would: signing in, navigating, updating a record, and confirming the result. No APIs required. No system replacement or integration project. Your team defines the workflow steps in natural language once. Then Browser Agent follows those steps directly in the browser to complete the task and update the customer. Learn more about Browser Agent from Jad Chamoun ➡️ https://lnkd.in/gwgEFRqy

Zendesk is thrilled to welcome the Forethought team to our family. We’re changing how AI agents reach customers — introducing Forethought AI Agents by Zendesk: best‑of‑breed, self‑improving agents that run natively in Zendesk and extend across any platform so teams can plug powerful AI into their existing stack without ripping and replacing. This accelerates our roadmap and delivers faster resolutions, sharper automation, and more choice for organizations adopting agentic service. Read the announcement: https://lnkd.in/gJ_BPnBt

Agentic AI isn’t just evolving—it’s moving toward full autonomy. In Dimitri Zabelin's part II of his PitchBook’s AI report on autonomous systems, our Head of AI, Antoine N., shares perspective on what this shift really means in practice. The takeaway: We’re moving beyond AI that responds → to systems that act. Agentic systems are increasingly able to plan, make decisions, and execute across workflows with minimal human input—marking a shift from assistive tools to operational drivers. But autonomy introduces new challenges as well. Such as, reliability and control, accountability, and real-world integration at scale. The companies that win won’t just adopt AI—they’ll design for it. Worth the read if you’re thinking about where AI goes next! 🔗 https://lnkd.in/ebt2YqtT

As promised, part II of my report on agentic AI with interviews from 13 different startups across six different industries, is officially published ⬇️ The industries they operate include: 🛅 Cybersecurity 💉 Health tech 📰 Legal tech 🧑💼 Enterprise SaaS 💳 Fintech 🤖 Robotics Across the interviews, the dominant undercurrent is that model capability is no longer the bottleneck. Governance, integration, and organizational readiness are. Adoption clusters in workflows that are structured, measurable, and reversible, while stalling where mistakes are costly or irreversible. A second theme is that trust is an engineering problem, earned through explainability, audit trails, confidence thresholds, and rollback safety, not just an organizational one. Finally, durable value is shifting from models to systems: defensibility now comes from proprietary context, deep workflow integration, data feedback loops, and platform-level control, with competitive dynamics starting to resemble ERP more than SaaS. Huge thank you to all of the contributors including Chris Sestito at HiddenLayer, Harvey, Andrew Miller at WRITER, Antoine N. at Forethought, 7AI, PathAI, BridgeWise, Shrivu Shankar at Abnormal AI, Mainul Mondal at Ellipsis Health, Jaspar Carmichael-Jack at Artisan, Skydio, Jazz, and Geordie AI. https://lnkd.in/g4KaYQSx

“Everyone is still talking about the best model. That’s not where the moat is anymore.” Al Martin, CX leader at Forethought talks about where AI is actually headed. As his Fast Company piece points out, the advantage is shifting to interoperability, or how well your AI connects across systems, data, and workflows. Because if everyone has access to the same models, differentiation disappears. What matters now - → Can your AI plug into the tools your business already runs on? → Can it take action across systems, not just answer questions? → Can it evolve without locking you into a single vendor? Interoperability isn’t just technical—it’s strategic. It’s what separates AI that sounds impressive from AI that actually drives outcomes for customers. Read more here ➡️ https://lnkd.in/gKn-jUan

Zendesk’s $500M acquisition sprint spans seven AI deals in three years to build a full stack from QA to self-improving agents. Over roughly three years, Zendesk made seven AI-focused acquisitions including Zendesk WFM (Tymeshift), Zendesk QA (Klaus), Zendesk AI agents (Ultimate), Local Measure, HyperArc (Acq. by Zendesk), Unleash (Acq. by Zendesk), and Forethought. Together, these additions cover workforce management, AI quality assurance, automation agents, voice, analytics, enterprise search, and self-improving agents. The latest deal, Forethought, is highlighted as a major milestone, with Zendesk’s CTO saying it accelerates their roadmap by more than a year. The total estimated spend across these acquisitions exceeds $500 million, with the stack progression outlined as QA to automation to voice to analytics to enterprise search to self-improving agents. Tom Eggemeier #AI #CustomerExperience #SaaS #Automation #EnterpriseAI

Forethought has built a strong foundation inside the conversation — resolving queries, retrieving knowledge, helping customers get to outcomes fast. But individual conversations aren't actually isolated from the rest of the business. Signals are constantly firing across a customer's ecosystem — billing events, product usage, dropped calls — and I've been thinking about what it would look like if CX teams had AI agents that could actually act on them. That's what we built Orchestrator to do. This isn't general-purpose workflow automation. Orchestrator is purpose-built for orchestrating AI agents across customer experiences — connecting triggers, business logic, and agents into workflows that operate reliably across every step of a CX process. It brings Forethought out of the conversation and into the broader customer operation. And customers are already running with it: → Automated ticket triage and routing (10k+ runs) → Outbound voice calls triggered by business signals (500+ runs) → Auto-creating cases for dropped calls (150+ runs) These aren't demos. These are real CX processes running today. We're just getting started — and I'm genuinely excited about where this is heading. The roadmap expands what's possible with agent orchestration significantly, and I think it changes what CX teams expect from their AI. https://lnkd.in/gD2359vZ