Field Engineer

Decagon launches AI for the era of personal agents at Dialogues. Search. New products for the next frontier headline the company's second annual flagship conference, plus new partnerships and industry-leading speakers Decagon, the leader in conversational AI agents for concierge customer experiences, today hosts its second annual flagship conference, Decagon Dialogues, ushering in an era of AI that meets customers - and their agents - however they choose to engage. With Voice 3, Personal Agent Gateway, Agent Modules, and Duet Apprentice, Decagon is meeting the moment where customers expect to be treated like someone the business already knows, not a ticket number in a queue. Frontier AI models are intelligent about the world in general, but a model that knows the world is not the same as one that knows a specific business: its policies, its systems, its customers, and the knowledge that used to live in one person's head. Closing that gap is what Decagon spent the last year building toward, and today's product launches are the result. Frontier models keep improving, but still don't account for the nuances of how an agent should sound on the phone, how it should interact with a customer versus that customer's personal agent, how a bank conducts a fraud dispute versus how an airline rebooks a stranded passenger, or how a company onboards and trains its best teammates. Each product announced today delivers that expertise: * Voice 3 knows customer experience, combining a new duplex architecture with Chord, the first voice model from Decagon Labs, for more natural customer conversations. Most voice agents combine general-purpose speech models with cascaded pipelines that struggle to coordinate conversation with ongoing work. Customers hear the result in stiff delivery and awkward pauses or interruptions. The Chord voice model is built for the intonation and natural pauses of a live call, slowing down for details like confirmation codes, then returning to a conversational pace. Voice 3 also introduces a new duplex architecture where conversation and task execution run in parallel, letting the voice agent narrate its progress and answer follow-up questions during a long-running task, rather than putting the customer on hold. * Personal Agent Gateway knows your customers however they arrive, helping you identify and serve personal agents on your terms.Personal agents like Muse, Instinct, and dots now book, buy, cancel, and negotiate for customers, and they are already showing up in support queues. Personal Agent Gateway detects whether a human or a personal agent is on the other end, using signals from your business, such as account history, and from the platform, like request cadence. Personal agents are routed to a dedicated channel where your agent resolves requests directly in a fraction of the time and cost. Personal Agent Gateway authenticates the agent, limits what it can do, and requires approval from the owner behind it for sensitive actions, declining if they can't be reached. * Agent Modules know your industry, extending your agent across the customer journey in financial services, travel and hospitality, healthcare, retail, technology, and telecom.Expanding beyond traditional support requires teaching your agent new workflows, requirements, and measures of success. Instead of requiring teams to build the infrastructure for each journey from scratch, modules bring together the orchestration each journey requires, adapted to your specific processes and business context. As modules extend your agent across more journeys, a shared management layer lets each team manage its part of the agent with separate access controls, while giving admins visibility across the whole customer relationship. Memory carries across conversations and channels, so the customer from a sales call last month is recognized when the onboarding call begins. * Duet Apprentice knows your business, enabling your agent to learn like your best people and keep pace as your business changes. With Apprentice, Duet studies your existing internal resources and learns from the escalated conversations your experienced reps handle. That knowledge grounds how it drafts Agent Operating Procedures (AOPs), runs Simulations, and analyzes conversations. Duet also keeps learning and working alongside your team after launch, following Slack and Microsoft Teams discussions as policies and plans change. Tag Duet into a thread and it can draft the AOP updates; when something is ambiguous, it proactively asks for clarity in the same channel. What Duet learns belongs to you, stays in your workspace, and isn't used to train models outside it. Frontier models know the world. They do not, on their own, know any one business, and that is the gap Decagon exists to close. "The promise of AI within customer experience is not just a support channel that happens to use AI, but a singular 'front door' for customers to interact with a business and carry them through the entire journey," said Jesse Zhang, Co-founder and CEO of Decagon. "With these new products and partnerships, we're bringing this vision to reality." Speakers across industries share how Decagon changed their businesses. This year's Dialogues lineup brings together experts from retail, healthcare, travel and hospitality, insurance, finance, and technology, all using Decagon to deliver concierge experiences. Select speakers include: * Eric Phillips, Senior Vice President and Chief Digital Officer, Delta Air Lines, joins the opening keynote to discuss how AI is reshaping customer experience at one of the world's largest airlines. * Samuel Liyanage, Vice President of Digital Customer Experience, and Misty Shafer Sterne, Vice President of Commercial Technology, both of American Airlines, explain how to design AI-powered journeys that build trust and loyalty. * Adam Terk, Vice President of Operations, Carrot, connects the healthcare experience across every touchpoint a patient has with the business and shares the outcome. * Diana Rainesh, Senior Vice President of Strategy & Analytics Support & Operations, Ticketmaster, describes using AI to power product discovery and personalization at scale. Partnerships take customer service to the next level In addition to recently announced partnerships with companies like OpenAI and AWS, Decagon continues to expand its ecosystem capabilities: * Decagon is partnering with Databricks to bring governed enterprise data into every customer conversation. Many of the most useful answers a Decagon agent could give - order status, entitlements, product usage, payment history - live in the Databricks Data + AI Platform rather than the conversation itself. Historically, surfacing that context meant building and maintaining bespoke data pipelines, with sensitive customer data copied into a new system each time. Working with Databricks, Decagon agents now read that record directly, so agents act on real-time customer context with no ETL and no new copy of the data. Conversation data flows back the same way, enriching Databricks tables with intent, sentiment, and root-cause insights. The partnership will also extend to inference: Decagon has joined Databricks' Built on program and will route model calls through Unity Gateway, so governance, lineage, and audit trails cover agent activity the same way they cover the rest of a customer's data estate. Decagon is also a launch partner on Databricks Marketplace, letting customers purchase Decagon through their existing Databricks relationship. * Decagon and Plaid are partnering so AI agents can complete complex financial tasks directly within a conversation. Many of the most common requests in financial services require the customer to log into their bank first. That means sending them outside the conversation to authenticate their account, which breaks the conversation and often ends it. Working with Plaid, Decagon agents can resolve account issues in real time, within the agent conversation. Consumers can complete a claim, recover a failed payment, manage applications, and handle disputes with full context from prior conversations, including the steps that require financial account access. The partnership ensures trust and safety, operating at the level of oversight financial services require. Plaid powers consumer-permissioned account access, and agent guardrails stay exactly as Decagon's financial services customers already run them.

Decagon says customers will be able to use OpenAI committed spending for its AI agents. The marketplace listing is still pending. Pricing, deployment options and instructions for applying commitments will follow. The audio brief. About 1:30 Read transcript Story brief. 3 key points For enterprise buyers, Decagon's planned OpenAI marketplace listing could make its customer-facing AI agents purchasable through existing OpenAI spending commitments, shifting the partnership from technical collaboration toward a procurement route. The arrangement does not change the agents' described model mix: Decagon combines OpenAI models with specialized models post-trained by Decagon Labs. The practical value depends on which customers qualify and how much commitment they can use - details Decagon has not yet disclosed. * 01 Decagon announced on October 2, 2026, that it is a launch partner for OpenAI's business-to-business marketplace. * 02 Only eligible OpenAI enterprise customers can apply part of a commitment; Decagon has not specified eligibility criteria or the amount available. * 03 Decagon says its marketplace listing will go live "soon," but has provided no launch date. An existing OpenAI spending commitment will give eligible enterprise customers another way to pay for customer-facing AI agents. Decagon said on October 2, 2026, that it is a launch partner for OpenAI's business-to-business marketplace. Under the arrangement, customers can apply part of their OpenAI commitment toward Decagon, although the company's marketplace listing is still to come. From technical partner to purchasing channel. The relationship predates the marketplace announcement. Decagon says it became a preferred partner in OpenAI's enterprise ecosystem earlier in 2026, working with the company on AI-powered customer experiences. The marketplace partnership builds on that earlier work. OpenAI models are an important part of Decagon's model lineup, the company says. They sit alongside specialized models that Decagon Labs post-trains - further trains for customer-experience work. That mixed lineup is the technical foundation Decagon describes behind its customer-facing agents, rather than a service built only around OpenAI models. Decagon's pitch for the new channel is simpler procurement: the purchasing process businesses use to approve and buy software. It argues that using the marketplace will reduce friction between a buying decision and deployment. That is the company's expected benefit of the arrangement. For businesses assessing the service, Decagon also points to configurable removal of personally identifiable information and layered guardrails that run before, during and after each conversation. These are safeguards the company describes for its agents, alongside its enterprise compliance claims. The next step is publication of the marketplace listing. Decagon says it will go live "soon," with deployment options, pricing and instructions for applying an OpenAI commitment. It has not given a specific launch date. In the meantime, interested businesses can contact its partnerships team at [email protected] to learn more about the planned offering. Sources. * decagon.aiDecagon is partnering with OpenAI on its new B2B marketplace | Decagon

AI & frontier Tech roundup - key developments in LLM agents, robotics, and model deployments. October 1, 2026 · openai/gpt-oss-120b TL;DR: The AI frontier is converging on three trends - richer multi-modal agent ecosystems, hardware-aware model scaling for robotics, and the release of higher-capacity, cost-effective LLMs that enable new product experiences. Multi-Modal agent tooling and deployments. * Karpathy urges a shift toward custom explainer artifacts - He recommends prompting LLMs to generate diagrams, HTML pages, or 3b1b-style videos, arguing that as models improve the bulk of work will become "discardable software artifacts" that require human oversight rather than manual creation @karpathy. * Cursor adds GLM 5.3 and GLM 5.3 Flash - Both the official Cursor announcement and a user note confirm the models are now available, with GLM 5.3 Max achieving the top score on CursorBench 4.0 @MiaAI_lab@cursor_ai. * TypeSafe AI showcases Jev for research agents - Jev delivers comparable report quality to a vanilla LLM judge while being 250x cheaper and 3-6x faster, and it avoids missed investigations that the baseline LLM suffered @typesafeai. * Tavus releases Griffin, the first video-Turing-test model - Griffin convinced 48 % of live participants it was human, a dramatic jump from sub-3 % pass rates, and it tops NVIDIA's full-duplex video benchmark @tavus. * NVIDIA open-sources 391 agent skills - The skill set spans CUDA, Jetson, robotics, and RAG, with benchmarks showing dramatic accuracy gains (e.g., Claude Code from 59 % to 97 % on the cuOpt skill) @undefinedKi. * DecagonAI launches a Personal Agent Gateway - The gateway lets enterprises identify and customize interactions with consumer agents such as Muse, Instinct, Dots, and Grok Bot, introducing the PACT protocol for provenance and permissioning @thejessezhang. * Ark builds a coordinated AI execution platform - Arkie AI App aims to orchestrate multiple agents into a single intelligent workflow, extending the single-agent paradigm @gizmocloud_ark. * Delvetown introduces a multi-agent society - Both the launch tweet and the research-lab description describe a shared world where human users and autonomous agents coexist, with early residents running on diverse models @lfschiavo@grove_research. Hardware-Aware model scaling for robotics and Physical AI. * Tesla trims onboard RAM for next-gen silicon - Elon Musk's team reduces RAM to 72 GB (AI5) and 144 GB (AI6) to alleviate supply-chain constraints while preserving bandwidth, relying on quantization (INT8/FP4) and tiered model loading to keep real-time performance @tslaming. * Boston Dynamics unveils high-DOF robot hands - The new hands feature 13 actuated degrees of freedom and are built for sim-to-real reinforcement learning, highlighting the need for precise actuation in AI-driven dexterity @BostonDynamics. * Axis Robotics partners with Manycore Tech - The collaboration combines physics-ready simulation assets with 3D generative models to improve spatial intelligence for Physical AI, emphasizing data-centric pipelines over raw compute @huynh_tinh1604. * Astribot T1 launches at IROS 2026 - Priced at $18 K, the cable-driven robot offers 23 DOF and an in-house Lumo-2 model, demonstrating a push toward affordable, high-fidelity manipulation platforms @XRoboHub. * Runway's Praxis-1 world-action model - Praxis-1 leverages large-scale video pre-training to learn embodied policies applicable across any robot embodiment, aiming to bridge the data scarcity gap in robot demonstrations @runwayml. * Research on context-aware agents - A new paper shows that letting strong models manage their own context improves performance by 11.4 % while cutting compute 21.5 % compared to rule-based summarization, a practical win for long-running agents @rohanpaul_ai. New high-capacity, cost-effective LLMs fueling frontier applications. * Gemini 4 Argon reaches human-level 3D understanding - The model tops Blueprint-Bench 2, suggesting near-human spatial reasoning that could accelerate robotics and biological simulations @DeryaTR_@ai_for_success. * Qwen 4B fine-tuned on scientific rewrites - Maxime Rivest reports that a 4 B-parameter model trained on 500 scientific publications already outperforms larger rivals, with fine-tuning completed in three hours on a 5-year-old GPU @MaximeRivest. * MiniMax H3 powers industry-specific video models - Boreal-H3 (advertising) and Utopai X (cinematic storytelling) illustrate how frontier video models can be specialized for niche domains while remaining open-weight @MiniMax_AI@ArtificialAnlys. * Local AI breakthroughs on consumer hardware - Users report running 35 B-parameter coding models on a RTX 3060 (12 GB VRAM) via aggressive quantization, achieving 262 K token context and competitive throughput, narrowing the gap between server-grade and edge deployments @Oluwaphilemon1. * GLM 5.3 adoption across ecosystems - Multiple tweets note rapid uptake of GLM 5.3 (including Flash) in Cursor, security research, and bug-hunting pipelines, indicating strong community momentum @MiaAI_lab@MiaAI_lab@guhe120. * Open-source agentic research loops - Hyperresearch demonstrates a 16-step Claude-Code pipeline that autonomously conducts deep research, verifies citations, and builds searchable vaults, showcasing the potential for fully automated scholarly workflows @beamnxw. Emerging business models around agentic finance and credit. * Agentics Credit proposes an on-chain credit score for agents - The system ties trading activity to a credit score, enabling agents to earn rewards and gain access to capital, moving beyond simple task-completion airdrops @kane_tdt. * Anvita Flow's TopNod Space aggregates market signals via multi-agent AI - The platform consolidates Reuters, Bloomberg, and other feeds into a unified UI, emphasizing organization over prediction for finance professionals @GesoraMeshack. * Grok Bot demonstrates autonomous profit generation - A user reports a self-funding Grok Bot that grew $70 to $8,340 in 16 hours by autonomously trading product-launch events, highlighting the feasibility of low-maintenance, revenue-generating agents @bl888m_eth. Community resources and infrastructure. * LLM-Tune.io integrates model selection, fine-tuning, agents, and security - The platform bundles the entire AI stack - from benchmark comparison to deployment audit trails - addressing the operational complexity of production AI @NgocO88803. * GitHub's Agentic Engineering System - A new internal tool aims to streamline agent development workflows, though details remain sparse @nick_cloudops. * OpenAI Dots architecture dissected - A community post outlines a ten-step blueprint for turning Dots into a persistent, cost-aware AI operating layer, emphasizing orchestration, memory sharing, and revenue loops @monokern. * e2e testing framework for agents - An open-source CLI enables deterministic and agentic API testing across web, mobile, and custom environments, supporting bring-your-own-agent pipelines @o_kwasniewski. Overall, the AI frontier is rapidly maturing: agents are becoming multimodal and interoperable, hardware-conscious model scaling is unlocking affordable robotics, and ever-larger, cheaper LLMs are democratizing frontier applications from finance to scientific research.