Forethought
by Forethought (a Zendesk company)
Self-improving AI agents for customer support, now sold as Forethought AI agents by Zendesk.
Forethought is a suite of agentic AI products for customer support teams: Solve resolves customer issues end-to-end across chat, email and voice, Triage classifies and routes tickets, Assist copilots human agents, and Discover surfaces knowledge gaps. Zendesk acquired the company in March 2026 and sells it to Zendesk and non-Zendesk customers alike.
Forethought builds AI agents that automate customer support work inside existing helpdesks. The product line has four agents: Solve, an omnichannel agent that resolves customer issues end-to-end across chat, email and voice; Triage, which classifies incoming tickets and routes them to the right queue or person; Assist, a copilot that suggests responses to human agents inside the helpdesk; and Discover, which analyses conversations to find gaps in the knowledge base. Its distinguishing idea is what Forethought and Zendesk call the Resolution Learning Loop: agents detect gaps in existing workflows, generate new procedures to close them, and test those changes before deploying, rather than waiting for an admin to script every path. Recent capabilities include native voice automation and computer-use style actions that operate legacy systems where no API exists. Forethought won TechCrunch Disrupt Battlefield in 2018, raised a reported $115 million from investors including NEA and Sound Ventures, and by 2025 supported more than one billion customer interactions a month for customers such as Upwork, Grammarly, Airtable and Datadog. Zendesk announced its acquisition on March 11, 2026, described by press as its largest deal in two decades, and closed it on March 26, 2026. Since June 4, 2026 the product has been sold as Forethought AI agents by Zendesk, purchasable as a Zendesk add-on or standalone by teams on other helpdesks. Pricing is quote-only and combines a platform fee with outcome-based charges.
A VP of Customer Support running a high-volume, ticket-rich helpdesk (Zendesk or Salesforce) who wants autonomous resolution rather than a deflection chatbot.
Routine tickets resolved end-to-end across chat, email and voice, with the agent proposing and testing new workflows itself instead of waiting on manual scripting.
At a Glance
- Category
- AI Agents & Orchestration
- Pricing
- Contact for pricing, Usage-based
- Target Market
- VPs of Customer Support, CX Operations Leaders, Contact Centre Leaders, CIOs
- Deployment
- Cloud-only
- Founded
- 2017
Key Features
- ✓Solve Agent
Omnichannel AI agent that resolves customer issues end-to-end across chat, email, voice and Slack instead of only deflecting to articles.
- ✓Resolution Learning Loop
Detects gaps in existing workflows, generates new procedures and tests them before deployment, reducing the manual retraining admins must do.
- ✓Triage Agent
Classifies incoming tickets by intent and routes them to the right team, cutting manual sorting and first-response time.
- ✓Assist Copilot
Suggests responses and relevant knowledge to human agents inside their helpdesk in real time, so escalated tickets close faster.
- ✓Discover Agent
Analyses support conversations to surface knowledge-base gaps and automation opportunities, telling teams which content or workflows to build next.
- ✓Voice and computer-use automation
Handles high-volume phone interactions natively and can operate legacy systems without APIs, extending automation beyond chat channels.
- ✓Automated PII redaction
Redacts PII, PHI and financial data during ingestion and deletes originals within 24 hours, easing regulated-industry security reviews.
Capabilities
Use Cases
- •Autonomous ticket resolution
A marketplace support team lets Solve resolve account and billing questions end-to-end, shrinking queues and time to resolution for customers.
- •Ticket routing at scale
An operations team uses Triage to classify every inbound email by intent and route it automatically, removing a manual sorting step before agents start work.
- •Voice call automation
A consumer brand automates repetitive phone requests such as order status with voice agents, freeing contact-centre staff for complex calls.
- •Knowledge-gap analysis
A knowledge manager uses Discover to find the recurring questions with no help article, then prioritises writing content that raises self-service rates.
- •Agent copilot for escalations
Human agents handling escalated tickets get suggested replies and relevant articles from Assist, shortening handle time on harder issues.
Ideal For
Best For
- ✓Enterprise support teams on Zendesk that want a native add-on for autonomous resolution
- ✓High-volume B2C support organisations with large historical ticket archives to learn from
- ✓Contact centres automating repetitive voice calls end-to-end
- ✓Support operations teams that need automated ticket classification and routing (Triage)
- ✓Knowledge managers looking for data on which help-centre articles are missing (Discover)
Not Ideal For
- ✗Small teams with little ticket history: third-party reviewers report the models work best with a large backlog of historical tickets and take weeks to learn
- ✗Buyers who need self-serve signup or published list prices: there is no free trial, only a sales-led proof of value, and every tier is quote-only
- ✗Teams wary of roadmap risk after an acquisition: the product is now owned by Zendesk, which some reviewers flag as uncertainty for customers on competing helpdesks
Deployment
Market Analysis
Pros
- ✓Mature product: TechCrunch Battlefield winner in 2018 and more than one billion customer interactions a month by 2025
- ✓Covers resolution, routing, agent assist and knowledge analytics in one suite
- ✓Strong compliance posture: SOC 2 Type II, ISO 27001, HIPAA-aligned controls and GDPR/CCPA DPA
- ✓Zendesk ownership brings a large distribution channel and a stated commitment to continuity for existing customers
Cons
- ✗Reviews cited by third parties report ongoing tuning of intents and workflows and inconsistent escalation to humans
- ✗Analytics dashboards are described by reviewers as messy, and some admin changes reportedly require vendor backend work
- ✗Quote-only pricing with no self-serve trial makes cost comparison slow
- ✗Post-acquisition product direction is uncertain for teams on non-Zendesk helpdesks
Pricing
Team
Contact for pricing
- ✓AI agents for chat and mobile
- ✓1 QA rubric
- ✓Up to 2 brands
Professional
Contact for pricing
- ✓Adds email, voice and Slack channels
- ✓More QA rubrics and brands
- ✓Custom handoff models
Enterprise
Contact for pricing
- ✓All channels
- ✓APIs
- ✓Advanced analytics
- ✓Up to 20 QA rubrics and brands
No list prices are published. Forethought charges a blend of a platform access fee and outcome-based pricing on resolutions, with add-ons for multibrand, an Analytics API and Discover; overage charges apply beyond plan limits. Instead of a free trial it runs a sales-led proof of value on your own data. Since June 2026 it is bought through Zendesk sales, as an add-on or standalone.
Security & Compliance
Sources
This page was written from 9 sources, 6 on domains other than forethought.ai.
- 1.forethought.ai — forethought.aivendor
- 2.forethought.ai — pricingvendor
- 3.forethought.ai — securityvendor
- 4.techcrunch.com — zendesk acquires agentic customer service startup forethough
- 5.zendesk.com — zendesk completes acquisition of forethought
- 6.support.zendesk.com — 10850639885082 Announcing Forethought AI agents by Zendesk f
- 7.cmswire.com — zendesk acquires forethought for self learning ai agents
- 8.thenextweb.com — zendesk acquires forethought agentic ai
- 9.eesel.ai — forethought review
Stay Ahead of the Curve
Weekly enterprise AI insights for technology leaders. No spam, no vendor pitches—unsubscribe anytime.
SubscribeRelated Products
Dust
Multiplayer AI agent platform where teams build, share and govern company-wide AI agents on their own data
SuperNinja Enterprise
24/7 AI employees on open-weight models, deployed in your own cloud for a fixed annual fee
Capacity
AI-native customer experience platform: omnichannel AI agents, real-time agent assist and automated QA on one knowledge layer
Relevance AI
No-code AI agent builder for sales, marketing, support and ops teams to build and run a multi-agent AI workforce