Agents AI

Forethought logo

Forethought

Autonomous AI for customer support

Customer Support
Visit website
From CustomForethought Technologies (acquired by Zendesk, March 2026)Founded 2018Reviewed Jun 2026

Our take

Our verdict

5.6/10

Enterprise AI platform for customer support that automates ticket resolution, routing, and agent assistance using LLMs.

Best for: Established enterprises handling 2,000+ support tickets/month that need autonomous resolution across multiple channels

Overall score5.6/10
Capability7.0
Ease of use4.0
Value for money4.0
Reliability6.0
Support & docs7.0

Pros

  • 70+ pre-built integrations covering Zendesk, Salesforce, Intercom, ServiceNow, and Freshdesk
  • True omnichannel reach across chat, email, voice, SMS, Slack, and mobile
  • Browser Agents (launched Oct 2025) act in browser-based systems without requiring API access
  • Intelligent triage and routing reduces manual ticket classification at scale
  • Published case studies show 65% average self-serve rate and 80% ticket deflection

Cons

  • No public pricing; annual contracts typically run $40K–$160K with usage-based overage fees
  • Requires 20,000+ historical tickets to perform well — prohibitive for smaller or newer teams
  • Setup takes 30–90 days; no free trial, no sandbox, and buying is sales-gated
  • Escalation logic is fragile — users report repetitive conversation loops instead of clean handoffs

Overview

Forethought is an enterprise AI platform purpose-built for customer support automation. Founded in San Francisco in 2018 — and a TechCrunch Battlefield winner that year — the company raised $115 million before Zendesk acquired it in March 2026. The product is now marketed as Forethought AI Agents by Zendesk, though it continues to operate as a standalone platform with its own pricing.

At its core, Forethought runs four product lines: Solve (autonomous resolution), Triage (routing and classification), Assist (live-agent copilot), and Discover (knowledge gap identification). The underlying technology is SupportGPT, built on OpenAI models combined with information-retrieval systems tuned on enterprise ticket data. The platform powers over 30 million customer interactions annually across 70+ integrations.

In practice, Forethought performs best for large, data-rich support organizations. It requires at least 20,000 historical tickets before its models generalize reliably, and onboarding typically takes 30 to 90 days. Pricing is entirely sales-led — Vendr data from 42 purchases shows a median annual spend near $60,000–$75,000, with contracts ranging from $40K to $160K plus per-deflection fees. There is no free tier and no self-service trial.

The Zendesk acquisition improves distribution but also narrows the platform's appeal outside Zendesk customers, and ongoing concerns about escalation loop failures and opaque reporting temper an otherwise capable product.

Key Benefits

  • Autonomous Resolution at Scale: Solve Agent handles chat, email, voice, SMS, and mobile tickets end-to-end, averaging 65% self-serve resolution rates in published case studies.
  • Deep Helpdesk Integrations: Pre-built connectors for Zendesk, Salesforce, Intercom, ServiceNow, HubSpot, and 65+ other platforms reduce implementation friction for enterprises already committed to those ecosystems.
  • No-API Browser Actions: Browser Agents (added October 2025) can operate in legacy browser-based systems that lack APIs, extending automation reach without custom development.
  • Agent Productivity Tools: Assist surfaces ticket summaries, suggested replies, and relevant KB articles in real time, measurably reducing average handle time for human agents.

Use Cases

  1. High-Volume Ticket Deflection — Deploy Solve Agent to handle FAQs and routine requests across all channels, freeing human agents for complex or sensitive issues.
  2. Intelligent Triage and Routing — Use Triage to classify incoming tickets by intent, sentiment, and language, then route them to the right queue or team without manual review.
  3. Agent Copilot — Equip frontline agents with Assist to surface context, draft replies, and recommend knowledge articles mid-conversation, reducing ramp time for new hires.
  4. Knowledge Base Maintenance — Run Discover continuously to flag unanswered question patterns and auto-generate draft KB articles, keeping documentation current without manual audits.
Customer Support AI
Ticket Automation
Agentic AI

Features

  • Solve Agent: autonomous ticket resolution across chat, email, voice, SMS, and mobile
  • Triage Agent: ML-based routing by intent, sentiment, language, and priority
  • Assist Agent: real-time suggestions, summaries, and KB article recommendations for human agents
  • Discover: automated knowledge-gap detection and article generation
  • Browser Agents: take actions in browser-based tools without API integration
  • Autoflows: natural-language workflow builder for complex branching logic
  • 70+ helpdesk, knowledge base, and CRM integrations
  • Multilingual support with custom triage models

Pricing

Basic
Custom (contact sales)
  • AI agent for chat only
  • Autoflows (natural language workflows)
  • Custom actions
  • Insights dashboard
  • Standard ticket triage models
Professional
Custom (contact sales)
  • Everything in Basic
  • Omnichannel AI agent: chat, email, voice, SMS
  • Multi-brand AI agents
  • Custom triage models
  • Multilingual support
Enterprise
Custom (contact sales)
  • Everything in Professional
  • Forethought API access
  • Knowledge base gap detection and article creation
  • Autoflow gap detection and generation
  • Advanced security and compliance
  • Premier support

Alternatives to Forethought