Hybrid AI is no longer just a concept for conference slides. It is already reshaping how contact centers, healthcare providers and global enterprises work at scale.
In a recent conversation with tech analyst and creator Evan Kirstel, Teneo AI’s CEO, Per Ottosson, shared how Teneo helps customers move from pilots to production, across millions of calls and dozens of languages.
This article is our point of view on that discussion and how it connects to the way we build Hybrid AI on Teneo.
From LLM Hype to Hybrid Reality
Per starts with a simple observation: large language models (LLMs) are transformative, but they are not enough on their own for enterprise CX.
Why Hybrid AI Matters
In real deployments, you need both sides of AI:
- Probabilistic AI for fluent, natural, long form conversation
- Deterministic AI for strict business rules, compliance and repeatable outcomes
What Hybrid AI Means on Teneo
Hybrid AI in Teneo combines our deterministic understanding and orchestration with best-in-class LLMs so enterprises get:
- Conversations that feel human and flexible
- Processes that are predictable and safe to put in front of customers
- Cost and latency that remain under control, even at huge volumes
Instead of “just throw an LLM at it”, Teneo provides a hybrid engine that can think, understand and act inside an enterprise grade framework. All while doing real LLM Orchestration.
Teneo results: 9.1 out of 10 Million Calls Automated
The most striking part of the conversation is not a roadmap. It is what is already live.
For a major “Mag 7” brand, Teneo 8 powers:
- 10 million inbound phone calls per month
- 9.1 million of those calls are fully automated
- Across 46 languages
Built once, expanded to 40+ languages over a weekend
A team of 5 people:
- Built the experience in English first
- Used Teneo’s predeveloped language resources (86+ languages) together with LLMs
- Deployed to more than +40 languages over a single weekend
- Spent roughly two weeks per language on fine tuning and optimization
The outcome is 91 percent automation. The remaining 9 percent of calls can be routed to human agents in one or a few locations, instead of standing up new contact centers in every country. You can find more on this in our case study.
For a US company expanding into new markets, that changes the operating model. You no longer need to open new offices, hire staff and provide training everywhere to support local customers in their language. Teneo 8 becomes the scalable backbone for global CX.
Inside Teneo 8: Key Capabilities Behind the Results
Per highlights several elements of the Teneo 8 release that make this level of automation possible.
AI Agent Builder
- Visual environment for defining goals, skills, tools and integrations
- Lets teams design complex journeys without wrestling with raw prompts
- Makes hybrid flows reusable across channels, brands and regions
Smarter Use of LLMs
- Not every interaction needs to hit an LLM
- Clear, structured intents are handled deterministically in Teneo
- LLMs are used where reasoning, summarization and free form language are needed, which reduces cost and latency
Public APIs
- Teneo can be used as an OEM, and live “behind the scenes” inside existing applications and contact center platforms (CCaaS)
- You build your own UI, workflows or products
- Teneo runs in the background as the automation engine for calls, chats and tasks
Teneo is no longer only the visible virtual assistant. It is also the invisible infrastructure that partners and providers can embed.
Meeting Customers Where They Are
Most enterprises already have IVR or bots in place. Teneo 8 is designed to plug in, not force a rip and replace.
Three Common Starting Points
- Legacy IVR: Replace 1980s like DTMF “press 1, press 2” menus and legacy systems like Nuance systems with conversational automation.
- Existing bot platforms: When Dialogflow or similar hits scaling limits, Teneo can sit on top or beside it, orchestrating complex journeys while the existing bot handles simpler flows.
- Cloud ecosystems: With integrations to Microsoft, Amazon and Google, enterprises can keep their CCaaS and cloud stack while Teneo powers the high value journeys.
You can layer Teneo into what you already have, then move critical journeys onto a hybrid AI foundation step by step.
Why So Many AI Projects Fail
Despite the hype, many conversational AI projects disappoint. This was highlighted in a recent study done by MIT where it was discovered that 95% of AI pilots never reached production.
Per sums it up simply: If it looks simple in an hour, it will probably fail in a month.
Beyond Lipstick on an LLM
Many tools just wrap a single prompt in a nice UI. They ignore:
- Edge cases and long-term maintenance
- Governance for HIPAA, SOC, ISO and similar standards
- Encryption, data residency and confidential compute
- State management, testing, analytics and deployment pipelines
Enterprises also need skilled teams, not just a weekend builder. Teneo’s hybrid approach accepts this complexity upfront, so solutions still work when you reach millions of interactions.
How To Start Your Hybrid AI Journey With Teneo
Per’s advice is straightforward. Follow this simple playbook:
- Pick one high value journey for your industry – For example billing, device support, policy changes or appointments.
- Define clear metrics – Containment, cost per contact, handle time, customer satisfaction and revenue.
- Apply hybrid AI deliberately – Let deterministic logic and LLMs each do what they are best at.
- Scale once it works – Reuse skills and integrations across new journeys, brands and channels.
Closing Thoughts
Hybrid AI is not a label for Teneo. It is the way we design, orchestrate and govern AI so enterprises can trust it in production.
Per’s conversation with Evan Kirstel gives a snapshot of that approach. The real proof is how it already runs today, from Medtronic’s virtual assistant to Mag 7 scale contact centers and the partners bringing hybrid AI to thousands of smaller businesses. Contact us to learn more!