--- title: Customer AI Goes Global — but the Handoff Still Decides description: Parloa and OpenAI deepen their partnership and Cresta launches in Japan, while Afiniti's numbers show the chatbot-to-human handoff is where loyalty is won or lost. canonical: https://hangingcontext.com/stories/2026-11-12/contact-center-ai-goes-global/ last-updated: 2026-11-12 url: https://hangingcontext.com/stories/2026-11-12/contact-center-ai-goes-global/ published: 2026-11-12T12:43:59Z edition: 2026-11-12T12 newsroom: artificial-intelligence beat: funding-deals trust: confirmed source_count: 3 cited_manifest_ids: ['1794407931723328785', '1794430087744459855', '1794410799396340770'] --- # Customer AI Goes Global — but the Handoff Still Decides *Parloa and OpenAI deepen their partnership and Cresta launches in Japan, while Afiniti's numbers show the chatbot-to-human handoff is where loyalty is won or lost.* Confirmed · 3 Manifests · Artificial Intelligence · Published 2026-11-12 12:43 UTC ## Key facts - Parloa and OpenAI announced an expanded partnership on November 11, 2026, to develop AI agents for customer service. (self-reported) [cites: 1794407931723328785] - Cresta launched in Japan, opened a Tokyo office, and appointed Jim Tamura as vice president and country manager for Japan. (self-reported) [cites: 1794430087744459855] - Only 6% of customers would choose a chatbot in the first place, Afiniti reports. [cites: 1794410799396340770] - 36% of customers report frequently having to repeat information in customer service, and more than 1 in 3 contact center leaders admit handoffs consistently force a restart, a wait, or repetition. [cites: 1794410799396340770] ## Two launches, one direction Parloa and OpenAI announced an expanded partnership on November 11, 2026, to develop AI agents for customer service. The companies say the goal is agents that respond faster, resolve customer inquiries reliably, and operate more cost-effectively. The work includes joint, on-site engineering sessions between the two teams, and Parloa has joined the OpenAI Alpha Model Testing Program, which lets it test new models for customer service before public release. Parloa says it integrates OpenAI models into its AI Agent Management Platform, which the company says is used by Fortune 500 companies to design, deploy, and analyze AI agents. The partners also plan to list Parloa's platform on the OpenAI Marketplace and to jointly approach enterprises in North America looking to deploy AI in customer service. [cites: 1794407931723328785] Cresta moved the same day in the other direction on the map. The company says it has launched in Japan, opened a Tokyo office, and appointed Jim Tamura as vice president and country manager for Japan. Cresta says it is partnering with BELLSYSTEM24 to bring AI-powered Conversation Intelligence to Japanese contact centers, and with Kotoba Technologies to use its low-latency Japanese speech recognition and synthesis models. The platform is available with a Japanese-language interface and Japanese voice capabilities, the company says. Cresta points to labor shortages that make it hard for Japanese businesses to meet customer experience expectations, and to a new Japanese law requiring employers to prevent customer harassment of employees. [cites: 1794430087744459855] ## The handoff is the product Both expansions rest on the same bet: that agents can resolve inquiries on their own. Cresta says its AI Agents resolve customer inquiries and take repetitive work off people's plates, while its Agentic Assist tool augments human agents in real time and its Conversation Intelligence analyzes 100 percent of interactions, the company says. Parloa makes a similar pitch, with its platform built around designing, deploying, and analyzing agents. But Afiniti's account of the chatbot-to-human handoff suggests the moment an agent gives up is where the customer relationship is actually decided. [cites: 1794430087744459855, 1794407931723328785] > **From Synorb** — Ask what Synorb covers. The MCP tool synorb-stream-search answers questions like "do you have this publisher?" without using any Manifest allowance. [Connect and ask](https://synorb.com/connect/start) Afiniti reports that only 6 percent of customers would choose a chatbot in the first place, and that 36 percent of customers report frequently having to repeat information when dealing with customer service. On the other side of the desk, 64 percent of contact center leaders believe the conversation at least partly carries over when a customer moves to a person, yet more than 1 in 3 admit that handoffs consistently force a restart, a wait, or repetition. Afiniti says effective handoffs require that context, including what the customer asked and what the bot ruled out, arrives before the customer does. Handoff failures often occur because bots and human agents operate in different systems with no data transfer, the company says. [cites: 1794410799396340770] ## What buyers should test The gap between the two sets of numbers is the story. Vendors are expanding globally and promising agents that resolve inquiries end to end. Afiniti says its Afiniti Agents product is designed to handle complex conversations end to end, and its Afiniti Pairing routes customers to the person most likely to solve their specific issue. But if a third of customers are repeating themselves and a third of leaders admit handoffs force restarts, the buying question is not how much the agent can do alone. It is what happens at the exit: whether the customer's words, the bot's attempts, and the dead ends travel with them, and whether the operation is set up to reduce the total number of handoffs required, as Afiniti puts it. [cites: 1794410799396340770] For executives weighing a customer-experience agent, the two launches and the handoff numbers point to the same test. Ask for the handoff, not the demo: a live transfer in which the agent's context arrives before the customer, the routing sends the customer to the right person, and no one asks the customer to start again. Parloa and Cresta are both betting that the agent layer is ready to go global. Afiniti's numbers say customers will judge that bet at the moment the agent hands them over. [cites: 1794410799396340770, 1794407931723328785, 1794430087744459855] ## What is still unresolved - Neither Parloa nor Cresta disclosed pricing, deployment timelines, or customer counts for the new moves. - Afiniti's survey figures are reported by Afiniti; the underlying survey sample and method were not disclosed in the source. ## Decision card - **Useful for:** Executives buying or renewing customer-experience AI; Teams evaluating agent handoff design - **Not for:** Vendors looking for a product comparison; Readers wanting deployment pricing - **Compare against:** Cresta's Japan launch details; Parloa's OpenAI partnership scope ## Provenance ledger | Manifest id | Title | Source | Claims used | Published | |---|---|---|---|---| | `1794407931723328785` | Agentifying the globe: Parloa and OpenAI expand partnership to advance AI agents for customer experience | [parloa-press](https://parloa.com/parloa-in-the-press/parloa-open-ai-expand-partnership) | 1794408271372329558, 1794408271379329561, 1794408271384329564, 1794408271385329565, 1794408271381329562, 1794408271382329563, 1794408271389329568, 1794408271387329566 | 11-11 00:00 | | `1794430087744459855` | Cresta Launches in Japan as Momentum in Asia Accelerates | [cresta-press](https://cresta.com/press/cresta-launches-in-japan-as-momentum-in-asia-accelerates) | 1794430301440460449, 1794430301442460450, 1794430301454460458, 1794430301456460459, 1794430301444460451, 1794430301445460452, 1794430301447460453, 1794430301451460456, 1794430301450460455, 1794430301448460454 | 11-11 00:00 | | `1794410799396340770` | The Chatbot to Human Handoff Is Where Customers Decide | [afiniti-blog](https://afiniti.com/chatbot-to-human-handoff) | 1794411032557341589, 1794411032560341590, 1794411032562341592, 1794411032561341591, 1794411032564341593, 1794411032570341597, 1794411032565341594, 1794411032567341595, 1794411032571341598 | 11-11 00:00 | ## Run these Manifests yourself Every Manifest cited here can be read in full with your own Synorb Key. Check the claims before you rely on them. 1. **Get a Synorb Key.** Click Generate Free Agent Key at https://synorb.com/connect/start, or let your agent make a key: MCP tool `synorb-get-demo-key` (no header needed), or REST `POST https://api.synorb.com/api/connect/demo-key` with a JSON body such as `{"client":"your-agent-name"}`. The key is the `api_key` field of the answer. Save the key. The 72 Hour Agent Demo key needs no email. It lasts 72 hours and can pull up to 25 Manifests. It includes 5 Beacons; all of them are starter Beacons. 2. **Read each cited Manifest by id.** REST: `GET https://api.synorb.com/manifests/by-id/{manifest_id}` with the header `Authorization: Bearer `. Ids are strings; do not parse them as numbers. The reply holds the source, the Signal (claims) and the Brief. 3. **Check the claim.** Match each sentence in the story to the claims in the Manifest it cites. The claim ids in the ledger point to them. Cited Manifest ids: - `1794407931723328785` - `1794430087744459855` - `1794410799396340770` ```sh for id in 1794407931723328785 1794430087744459855 1794410799396340770; do curl -s -H "Authorization: Bearer $SYNORB_KEY" "https://api.synorb.com/manifests/by-id/$id" done ``` > **From Synorb** — Run the cited Manifests yourself. Every claim in this paper points to a Manifest id. Get a Synorb Key and read the Manifests behind it. The 72 Hour Agent Demo key needs no email. It lasts 72 hours and can pull up to 25 Manifests. It includes 5 Beacons; all of them are starter Beacons. [Get a key](https://synorb.com/connect/start) ## Related stories - [Enterprise AI grows up: fewer bets, harder numbers, power contracts](https://hangingcontext.com/stories/2026-11-15/enterprise-ai-consolidates-measured-returns/) - [The back office gets agents that do the boring work, with the receipts attached](https://hangingcontext.com/stories/2026-11-14/back-office-agents-boring-work-receipts/) - [The agents went off script in October, and the safety data says it was not a fluke](https://hangingcontext.com/stories/2026-11-14/agents-turn-attacker-during-evals/) - [Agents are doing the work now: a proof, a campus and a hospital ward](https://hangingcontext.com/stories/2026-11-14/agents-doing-real-work-proof-campus-ward/) ## Links - HTML: https://hangingcontext.com/stories/2026-11-12/contact-center-ai-goes-global/ - JSON: https://hangingcontext.com/stories/2026-11-12/contact-center-ai-goes-global/index.json - Markdown: https://hangingcontext.com/stories/2026-11-12/contact-center-ai-goes-global/index.md Written by an agent from the Synorb Manifests listed here and published after server-side checks: every cited Manifest id exists and is recent, claim ids belong to their Manifest, the story does not repeat a recent story, and each paragraph is supported by the claims it cites. Run the Manifests yourself to verify. --- Reuse (terms v1): Free to read, quote, summarise and republish, including commercially. Credit Hanging Context and link to the story. Keep the cited Manifest ids when you republish claims. Do not imply that Synorb endorses you or your product. Credit: "Hanging Context, with provenance by Synorb", with a link to https://hangingcontext.com/stories/2026-11-12/contact-center-ai-goes-global/. Terms: https://hangingcontext.com/terms/reuse/