> ## Content Index > Fetch the complete content index at: https://www.theaidirectory.com.au/llms.txt > Use this file to discover other available public pages before exploring further. # Forethought vs Decagon (2026): Proven Deflection or Autonomous Resolution? - URL: https://www.theaidirectory.com.au/compare-forethought-vs-decagon-australia/ - Published: 2026-06-27T23:00:00.000Z - Updated: 2026-10-09T06:40:46.000Z - Description: Forethought is the established deflection-and-triage layer. Decagon is the agentic challenger pitching higher full-resolution rates. Which one suits an Australian CX team? - Author: Adam - Tags: Customer Service & AI Chatbots, Support, #comparison Head-to-head verdict [**13**/25ForethoughtWinner](https://www.theaidirectory.com.au/review-forethought-australia-2026/)[**14**/25Decagon](https://www.theaidirectory.com.au/review-decagon-australia-2026/) **Our pick: Forethought.** The verdict below says who should buy the other. > **Verdict (TL;DR):** Two US enterprise support platforms, two different bets. Forethought (our score: 13/25, ⭐⭐⭐ Good) is the established layer that sits on your helpdesk and deflects and triages tickets. Decagon (our score: 14/25, ⭐⭐⭐ Good) is the newer agentic challenger that pitches higher end-to-end resolution and self-improving agents. Neither bills in AUD, neither confirms Australian data residency, and neither publishes a price. For a mid-market Australian CX team that wants proven deflection on an existing stack, **Forethought is the safer pick**. For a high-volume operation chasing the highest possible full-resolution rate — and willing to run a heavyweight procurement — Decagon earns the look. Most Australian SMBs should buy neither. > > *Last reviewed: 28 June 2026 by Adam.* ## The tension This is not a small-business decision. Both tools are enterprise infrastructure, sold through a sales team, priced in US dollars, with nothing on the pricing page but a "Get a Quote" or "Get a demo" button. If you run a five-person support desk on a Shopify store, neither belongs on your shortlist. Start with our [best AI customer-service chatbots for Australia](https://www.theaidirectory.com.au/best-ai-customer-service-chatbots-australia/) roundup instead, where the tools have AUD prices and free trials. For everyone still reading — the contact-centre managers, the CX operations leads — the choice comes down to a strategic bet. Forethought is the incumbent layer. It bolts onto Zendesk, Salesforce, Intercom and around 70 other helpdesks and does three jobs well: deflect what it can, triage what it can't, and arm your human agents with summaries and drafts. It has a track record. It is a known quantity. Decagon is the agentic challenger. It was built the other way around — the AI agent is the primary operator, and humans are the exception, not the rule. It publishes the best-documented resolution numbers in the category and, in June 2026, shipped a self-improving "Duet Autopilot" agent. It is newer, more ambitious, and pitched at higher autonomy. Proven deflection versus newer autonomous resolution. That is the real fork. ## What they cost in Australia Neither vendor publishes a dollar figure. Both are quote-only, USD-only, and sales-led. | | Forethought | Decagon | | ---------------- | ------------------------------------------------------- | ----------------------------------------- | | Published price | None — every tier is "Get a Quote" | None — only "Get a demo" | | Pricing model | Platform access fee + outcome-based cost per resolution | Not disclosed; annual enterprise contract | | Billing currency | USD | USD | | AUD billing | No | No | | Free trial | No — runs a "Proof of Value" pilot instead | No — no self-serve entry point | | Entry point | Sales process | Sales process | Sources: [forethought.ai/pricing](https://forethought.ai/pricing?ref=theaidirectory.com.au) and [decagon.ai](https://decagon.ai/?ref=theaidirectory.com.au), both checked 28 June 2026\. Forethought's pricing page states the model is "a blend of platform access fees and an outcome-based pricing cost." Decagon's site shows no pricing of any kind — the only call to action is "Get a demo." — [Wayback 29 June 2026](https://web.archive.org/web/20260629235915/https://decagon.ai/) **What you'll actually pay:** Because neither vendor discloses figures, the only numbers available are second-hand, and we treat them as budgeting ballpark, not quotes. For Forethought, third-party procurement marketplaces — not Forethought — put typical annual contracts at US$40,000–US$160,000, with a reported median near US$59,500\. At that midpoint that's roughly US$59,500/yr, about A$83,300 ex GST at US$1 ≈ A$1.40\. For Decagon, there is no comparable published range; its customer roster (Duolingo, Hertz, Rippling, Affirm) points to large enterprise contracts, but any specific figure would be a guess we won't print. Assume both sit firmly in five- or six-figure annual territory. **GST:** Both are USD imported services. You'll likely handle GST as a reverse-charge on an imported service, and your final cost moves with the exchange rate on a contract you've already signed. Confirm the tax treatment with your accountant before signing either. ## Feature diff Both platforms do agentic, action-taking support — refunds, rescheduling, account updates — not just FAQ deflection. The difference is emphasis and maturity. | Capability | Forethought | Decagon | | ----------------- | -------------------------------------------------- | ------------------------------------------------- | | Core bet | Deflect + triage on your existing helpdesk | Autonomous end-to-end resolution | | Architecture | Four agents: Solve, Triage, Assist, Discover | AI agent as primary operator, human on escalation | | No-code workflows | Autoflows (plain-language policy logic) | Agent Operating Procedures (plain-language) | | Channels | Chat, email, voice, SMS, Slack | Voice, chat, email with shared context layer | | Agent copilot | Assist — summaries and drafts for humans | Available; less central to the pitch | | Routing / triage | Triage — sentiment, language, urgency tagging | Built in, less of a headline feature | | Self-improvement | Discover surfaces knowledge gaps | Duet Autopilot — autonomously proposes fixes | | Always-on QA | AI QA rubrics | Watchtower — every conversation checked | | Integrations | \~70, incl. Zendesk, Salesforce, Intercom, Gorgias | Salesforce, Zendesk, Shopify, Stripe, Intercom | The honest read: Forethought's strength is breadth and fit. Four connected jobs, \~70 integrations, and a copilot layer that helps your existing human agents rather than trying to replace them. It is the better choice if your goal is to add AI on top of a team you're keeping. Decagon's strength is depth of autonomy. The Agent Operating Procedures approach and the omnichannel memory layer are genuinely modern, and Duet Autopilot — a self-improving agent that proposes its own knowledge corrections — is a real step beyond static workflow builders. Its published case studies (80% deflection at Duolingo, 70% chat and voice resolution at Chime) are named and attributed, which is rare in this category. We haven't independently verified those numbers, but they aren't anonymous. Neither claim should be read as your result. Both vendors publish best-case figures. Your actual resolution rate depends on how clean your knowledge base is and how repetitive your tickets are. ## Australian fit This is our defining criterion, and it's where both tools struggle. We scored Forethought 2/5 on Australian Fit and Decagon 1/5\. Neither clears the bar to be named a Top Pick for Australians, and that gate governs the verdict below. **Data residency.** Neither vendor confirms Australian hosting. Forethought leans on default sensitive-data redaction, data-transfer agreements and international security accreditations (SOC 2 Type II, HIPAA, GDPR, CCPA) rather than in-country storage. Decagon enforces zero-day retention with its LLM providers, redacts PII automatically, and carries SOC 2 Type 2, but its security page names no specific regions and makes no mention of Australian or APAC residency. For AU government, health or financial-services buyers with data-sovereignty rules, that absence is a real obstacle for both. If local residency is mandatory, you'd need to confirm the position directly with each vendor. On the redaction-and-compliance posture, Forethought has a marginally more documented story for Australian regulated buyers — but "documented compliance" is not the same as "data hosted in Australia." Neither clears that bar publicly. **Local support and presence.** Neither names an Australian office, an APAC sales team, or AEST-hours support coverage. Plan around US-timezone support for both. For enterprise procurement, the absence of a local entity is a negotiation item, not necessarily a deal-breaker — but it is friction either way. **GST and currency.** Same story for both: USD billing, no AUD option, GST likely handled as a reverse-charge import. Neither publishes AU-specific tax documentation. **AU case studies.** Neither has published Australian reference customers. Decagon's case studies are strong but entirely US-based; Forethought's are global but not AU-specific. For Legal, Finance and HR-adjacent deployments, note that neither tool replaces qualified professional advice — confirm GST treatment and any data-sovereignty obligations with your accountant or legal adviser before signing a USD enterprise contract. ## Who buys which **Buy Forethought if** you're a mid-market Australian CX team running an established helpdesk — Zendesk, Salesforce, Intercom — and you want to add a proven deflection-and-triage layer without ripping out your stack or your people. Forethought's four-agent model assumes you're keeping your human agents and making them faster. Its breadth of integrations and its copilot layer make it the lower-risk way to get AI into an existing operation. The "Proof of Value" pilot lets you see results on your real data before committing. This is the safer bet for the team that wants AI on top of what it already runs. **Buy Decagon if** you're a high-volume operation — fintech, retail, telco, large health — chasing the highest possible full-resolution rate, and you're prepared to treat the AI agent as the primary operator rather than a helper. Decagon's autonomy ceiling is higher, its case-study numbers are the best-documented in the category, and Duet Autopilot's self-improvement is a genuine operational edge at thousands of SKUs or frequent policy changes. You'll pay for it in implementation complexity and a heavyweight procurement, and you'll need a CX operations team to run it. But if maximum autonomous resolution is the goal, Decagon is the more ambitious instrument. **Buy neither if** you're an Australian small or medium business. There's no self-serve path, no published price, no free trial, and no AUD billing on either. Go to the roundup, pick something transparent and AUD-friendly like Intercom Fin or Freshdesk Freddy AI, and graduate to a platform like these only when your volume genuinely demands it. ## The winner For the typical Australian buyer who lands on this comparison — a mid-market CX team that wants proven results without betting the whole operation on bleeding-edge autonomy — **Forethought is the pick.** It is the lower-risk way to add AI deflection and triage to a helpdesk you're keeping, its integration breadth fits common Australian stacks, and its Proof-of-Value pilot lets you test before you commit. On our scores it edges in on Australian Fit (2/5 to Decagon's 1/5), which is the criterion that matters most here. That is a narrow win, and an honest one. Decagon scores one point higher overall (14/25 to 13/25) on the strength of its feature depth and reliability posture, and for a high-volume operation specifically chasing the highest full-resolution rate, it is the better-aimed tool. But both fail our Australian Fit gate, both bill in USD with no local data residency, and neither can be named an unqualified Top Pick for Australians. Whichever you shortlist, run the pilot, confirm data sovereignty in writing, and get your accountant across the GST treatment before you sign. ## Frequently asked questions ### Is Forethought or Decagon better for an Australian business? For a mid-market Australian CX team wanting proven deflection and triage on an existing helpdesk, Forethought is the safer pick and scores marginally better on Australian Fit. For a high-volume operation chasing maximum autonomous resolution, Decagon is better aimed. Most Australian SMBs should choose neither — both are enterprise-only, USD, and sales-led. ### Do Forethought or Decagon publish pricing? No. Both are quote-only. Forethought's pricing page shows three tiers, all "Get a Quote," on a model that blends platform access fees with outcome-based costs. Decagon publishes no pricing at all — the only call to action is "Get a demo." Both bill in USD with no AUD option. ### Do either offer Australian data residency? Neither confirms Australian hosting publicly. Both rely on compliance certifications and data-handling controls rather than in-country storage. For AU government, health or finance buyers with data-sovereignty rules, confirm the residency position directly with each vendor before proceeding. ### Which has higher resolution rates? Decagon publishes the best-documented, named case-study numbers in the category — 80% deflection at Duolingo, 70% chat and voice resolution at Chime — and positions itself for the highest full-resolution rate. Forethought publishes strong figures too. Both are best-case vendor numbers; your actual result depends on your knowledge base and ticket mix. ## Read next - [Forethought review](https://www.theaidirectory.com.au/review-forethought-australia-2026/) - [Decagon review](https://www.theaidirectory.com.au/review-decagon-australia-2026/) - [Best AI customer-service chatbots in Australia →](https://www.theaidirectory.com.au/best-ai-customer-service-chatbots-australia/) - [How we score — our review methodology](https://www.theaidirectory.com.au/methodology/) --- ## Sources Primary (vendor) sources, checked 28 June 2026: - Forethought pricing model and tiers — [forethought.ai/pricing](https://forethought.ai/pricing?ref=theaidirectory.com.au) — Wayback: https://web.archive.org/web/20260621085754/https://forethought.ai/pricing - Decagon product, channels and pricing absence — [decagon.ai](https://decagon.ai/?ref=theaidirectory.com.au) — Wayback: https://web.archive.org/web/20260629235915/https://decagon.ai/ Existing AID reviews cited for scores and verdicts (not re-scored here): - Forethought review (13/25, ⭐⭐⭐ Good; Australian Fit 2/5) — review-forethought-australia-2026 - Decagon review (14/25, ⭐⭐⭐ Good; Australian Fit 1/5) — review-decagon-australia-2026 Forethought pricing-range figures (US$40k–US$160k, \~US$59,500 median) are third-party procurement-marketplace estimates, not vendor-published, and the body is explicit about that. No comparable published range exists for Decagon, so none is printed. Exchange rate used: US$1 ≈ A$1.40 (May 2026 basis). GST on both is likely a reverse-charge import — confirm with your accountant. --- *The AI Directory Australia is an independent editorial publication. We review and score AI tools for the Australian market and take no payment for positive coverage. Affiliate links, where present, are disclosed and never affect our verdict. Every review carries a last-reviewed date; flag anything that looks stale.*