Breaking
17,000+ AI agents live in enterprise contact centers TENEO trades on Nasdaq First North, Stockholm Telefonica Germany routes 900,000 monthly calls through Teneo Swisscom: 9M calls a year, +18 NPS Medtronic deployment: $6M saved, 36,000 agent hours freed Cost per call: $0.40 AI vs $5.60 live agent 86+ languages from a single instance 2001 founded as Artificial Solutions, rebranded 2024 17,000+ AI agents live in enterprise contact centers TENEO trades on Nasdaq First North, Stockholm Telefonica Germany routes 900,000 monthly calls through Teneo Swisscom: 9M calls a year, +18 NPS Medtronic deployment: $6M saved, 36,000 agent hours freed Cost per call: $0.40 AI vs $5.60 live agent 86+ languages from a single instance 2001 founded as Artificial Solutions, rebranded 2024
Company Profile / Enterprise AI / Stockholm

The 24-Year-Old Startup Quietly Answering Your Bank's Phone

AI that doesn't guess. Ever.

Call your telecom provider on a bad day and you might spend nine minutes on hold, then explain your problem three times to three different people. Now imagine the first voice that answers already knows why you called, speaks your language, and settles the matter before a human is ever paged. For a growing list of large companies, that first voice is built by a Swedish firm most people have never heard of: Teneo.ai.

Teneo runs software that automates customer service inside contact centers. Its AI agents pick up calls, read chat messages, and work through the routine questions that clog support lines - password resets, billing queries, plan changes, order status. The company reports more than 17,000 of these agents in production across global enterprises, handling the kind of volume that would otherwise require call floors full of people.

What makes the story unusual is the age. Teneo.ai is not a 2023 vintage AI startup. It was founded in Stockholm in 2001, under the name Artificial Solutions, and spent two decades in the unglamorous business of teaching machines to understand human language before the current wave made that skill fashionable.

17,000+
AI agents in production
86+
languages, one instance
99%
claimed intent accuracy
2001
year founded

01 / What it doesThe phone line, rebuilt

The product enterprises buy is the Teneo Platform, first launched in 2010. It is a development environment for building conversational AI, and it does the parts customers rarely see: understanding what a caller actually means, holding the thread of a conversation, pulling the right account data, and taking an action. Teneo bundles this into packages with plain names - Voice AI, Contact Center Automation, Agentic AI, and an Agent Copilot that whispers suggestions to human staff while they talk.

The company's own numbers describe the appeal. It says a single enterprise customer can save up to $32.4 million a month, that automation lifts customer satisfaction scores, and that roughly 60% of interactions get resolved without ever reaching a person. The tagline it now leads with is blunt: AI that doesn't guess.

"Build autonomous agents that reason, act, and resolve issues with 99% accuracy." Teneo.ai, company homepage

In practice, that means a company can point Teneo at the tasks it already knows are repetitive and let the software absorb them. A telecom might automate SIM swaps and outage checks; a retailer might hand over order tracking and returns; a healthcare firm might route device questions to an agent that never sleeps and never loses patience. The work is not glamorous, which is precisely why it is valuable - these are the calls that burn out human staff and pile up during a bad week.

02 / Who pays for itTelecoms, and their very long call queues

Teneo sells to companies with enormous inbound volume, where a percentage point of automation is real money. Its named customers cluster in telecom, healthcare, and retail: Telefonica, Swisscom, AT&T, Medtronic, HelloFresh, Circle K, and the Swedish insurer Folksam. These are not businesses looking for a chatbot demo. They are businesses looking to take pressure off call centers that never stop ringing.

The scale claims are specific. Teneo says Telefonica Germany routes around 900,000 calls a month through its voice agents. Swisscom, it says, runs roughly 9 million calls a year on the platform and lifted its Net Promoter Score by 18 points. A Medtronic deployment is cited at $6 million saved and 36,000 agent hours freed.

9M CALLS/YR TENEO VOICE AI answers first +18 NPS
The Swisscom line, drawn as a diagram: nine million calls a year that mostly never meet a human, and a satisfaction score that went up anyway.

03 / The problemWhy a call still costs $5.60

The economics are the whole pitch. A live agent handling a call is expensive - staffing, training, turnover, overflow during a product outage. Teneo puts the cost of a human-answered call at $5.60 and the cost of one its AI resolves at $0.40. Multiply that gap across millions of interactions and the finance department starts paying attention, which is exactly the department Teneo is built to reach.

Cost per resolved call, as claimed by Teneo
Live agent
$5.60
Teneo AI
$0.40
The bars are drawn to scale. That thin sliver on the bottom is the argument the company makes to every CFO it meets.

But cost is only half the problem. The other half is trust. A cheap AI that gives wrong answers is worse than an expensive human who gives right ones - it erodes the brand with every call. Teneo's positioning leans hard on this anxiety. The phrase "doesn't guess" is aimed squarely at the reputation that generative chatbots earned for confidently inventing things. Teneo wants buyers to believe its agents stay inside the lines.

04 / The differenceRefusing to marry one model

Most AI vendors tie you to a stack. Teneo's distinguishing bet is the opposite: it describes itself as LLM-agnostic, orchestrating multiple large language models rather than committing to one, and wrapping them in what it calls "100% output control." Underneath sits older machinery - a proprietary system called the Teneo Linguistic Modeling Language, which the company has used for years to parse intent across languages.

The practical claim is that an enterprise can swap the underlying model as the market shifts, keep its conversational logic intact, and constrain what the AI is allowed to say. For a bank or a telecom, that control is not a feature. It is the reason legal signs off.

Control is the product. The model is a component you can replace - the guardrails are what you buy.
Where the calls go
60%resolved by AI
Resolved without a human
Escalated to a live agent
Teneo's stated containment rate
Six in ten interactions never reach a person, by Teneo's account. The remaining four are handed up with the context already gathered.

05 / The long roadFrom a prize-winning chatbot to the call floor

Teneo's history is longer than its current branding suggests. In 2008, a chatbot from the company called Elbot won the Loebner Prize, an award for the most convincingly human machine conversation of the year. In 2013 it shipped a mobile assistant named Lyra, arriving before phones came with voice assistants baked in. In 2018 it raised 13.7 million euros. In 2019 it went public on Nasdaq First North through a reverse takeover, trading first as ASAI.

The pivotal move came in August 2024, when Artificial Solutions took the name of its own platform and became Teneo.ai, with the ticker changing to TENEO. The rebrand read as cosmetic and was not. It marked a company deciding that the words "voice" and "agentic" would define its next decade, and putting its public identity behind that call.

The founding trio - Johan Ahlund, Johan Gustavsson, and Michael Soderstrom - built the early company, and leadership has since passed to a management team that took it public and steered the rebrand. What has stayed constant across the leadership changes is the obsession with language: the belief that the hard part of customer service AI is not the talking, but the understanding.

2001
Founded in Stockholm
Three founders start the company that becomes Artificial Solutions.
2008
Elbot wins the Loebner Prize
A chatbot earns recognition for human-like conversation.
2010
Teneo Platform launches
The conversational AI environment that still carries the name arrives.
2019
Goes public
Lists on Nasdaq First North via a reverse takeover of Indentive AB.
2024
Becomes Teneo.ai
The rebrand repositions the company around voice and agentic AI.

06 / The businessSelling minutes, not licenses

Teneo makes money the way enterprise software usually does, with a twist for the AI era. It licenses the platform to large organizations on subscription and usage-based terms, then adds the implementation and professional services that a serious contact-center rollout demands. The usage part matters: because pricing tends to follow automated interactions, the vendor grows when the customer's automation grows, which aligns both sides toward pushing more calls through the machine.

The expertise behind the sale is not a slide deck. It is roughly a quarter century of computational linguistics, a team spread across Stockholm, Barcelona, and Austin, and a patent portfolio the company valued at over 1.6 billion Swedish kronor in 2022. That depth is the quiet reason a risk-averse bank or telecom is willing to put a Swedish small-cap in front of its customers - the company was solving natural-language problems long before the tools to solve them cheaply existed.

There is also a cultural tell in how Teneo talks. Where many AI firms sell wonder, Teneo sells restraint. Its marketing is thick with words like accuracy, security, and control, and thin on the promise that the software will surprise you. That is a deliberate read of its buyer. Nobody running a Fortune 500 support line wants a creative AI. They want one that does the same correct thing a million times in a row.

07 / The marketA small-cap in a crowded room

Teneo occupies a contested patch. On one side sit the cloud giants - Google's Contact Center AI, Amazon's Connect and Lex, Microsoft's Copilot Studio. On another sit focused challengers like Cognigy, Kore.ai, and the newer Sierra. Teneo's answer to all of them is the same: two decades of linguistics, a public balance sheet, and a promise of neutrality between the models everyone else is trying to sell you.

It is a publicly traded pure-play in a field where most rivals are either giant conglomerates or venture-funded newcomers. That makes it easy to overlook and, for the enterprises it serves, easy to underestimate. The work Teneo does is invisible by design. If it does its job, you never notice the software that answered - you just notice you got off the phone faster than you expected.

#conversational-ai#agentic-ai#voice-ai #contact-center#customer-service#llm-orchestration #enterprise-saas#stockholm#nasdaq#nlu