Leena AI unveils conversational AI ‘colleagues’ for the enterprise

Imagine driving to a business meeting while conversing with a voice-enabled AI agent, and asking it to open a lead for a prospect in Salesforce. Or strolling in the park, asking your AI assistant whether you have enough laptops in inventory, and if not, to order what you need. Or even, while waiting at the airport, asking the AI agent to put in a request on your behalf for personal time off (PTO).

That is the promise of voice AI, according to startup Leena AI, which Thursday launched agentic AI “virtual colleagues” that can actively listen and speak in natural language. Workers get AI assistance simply by talking, anytime and anywhere, via mobile and desktop apps and through Slack and Teams integrations.

“If you think about the future of AI in the enterprise, collaboration is a key, and collaboration has to happen over not just text, but voice,” Leena CEO and co-founder Adit Jain told Computerworld.

‘Extremely personified’ AI connecting employees and systems

Leena’s agentic AI colleagues can handle domain-specific and cross-domain requests, serving as go-betweens for employees and enterprise systems. However, the human is always in the loop; the agent must request approval before taking action.

Powered by Gemini and two orchestrator large language models (LLMs) fine-tuned on GPT-4.1, with ServiceNow, Workday, SAP, Oracle, ADP, Salesforce, and other integrations, the AI colleagues understand context and past conversations, and update themselves regularly to provide human-like discussions.

In one product demo, a worker sitting in a busy lobby interacted with Leena on his mobile phone, saying “Hi Leena, I just spilled a drink on my laptop. Can you order me another one?”

Leena, in a female voice, advised him to shut down his laptop and unplug it to stop further damage, then told him he needed to fill out a quick form to have a replacement ordered, and dropped a link into the chat. Leena then told him the appropriate department would handle things from there, and signed off, saying “Need help after that? Just give me a shout.”

Jain said a big debate within the company has been around how much personification the models should have; the consensus was the more the better, so Leena’s AI colleagues are “extremely personified,” with names, email addresses, phone numbers, and Slack and Teams profiles. Some even have favorite sports teams.

Jain noted that the more human and conversational an AI system is, the more quickly it will be adopted. This is particularly important in enterprises that need AI tools “because the applications are so broken, the experience is so bad, systems don’t talk to each other, data is everywhere.”

A ‘two-way street’

The company initially began working with text-based AI, but quickly realized the importance of the conversational element. Jain pointed out that 30% to 35% of usage among early customers is via voice. There is particularly high usage during transit times, as well as “high excitement” among field employees, salespeople, plant workers, and retail store employees, Jain noted.

Use cases span IT, HR, finance, marketing, sales, and procurement, Jain said, noting that the company serves customers across industries including finance, manufacturing, pharmaceuticals, and consumer and packaged goods. Jain said one chief experience officer (CXO) at a major industrial company predicted that incorporating voice-enabled AI colleagues could boost the productivity of IT and HR teams by 50%.

The goal is to ultimately make the AI colleagues even more autonomous. Jain imagines a future where an agent can be added to a Zoom, Slack, or Teams call not just to take notes, but to collaborate, exchange ideas and “actually do multiple layers of work.” In future releases, Leena’s AI will even be able to see users’ screens to assist with these tasks. 

But, Jain emphasized: “It’s a two way street: It’s not just there for me, it can also reach out to me for help.” Agents will be able to make and receive calls, send and receive texts, and proactively work alongside humans. “They can just ping you and say ‘Hey, how do I handle this?’ And you tell them, and they go handle it for you.”

Voice AI is ‘closing the uncanny valley gap’

Voice AI has advanced significantly in the past year, with Leena competing in a crowded space with the likes of Workai, Bloomfire, Elium, Sociabble, Glean, Assembly, and others.

“We’ve seen the power of this type of voice experience when paired with LLM interaction,” said Brian Jackson, principal research director at Info-Tech Research Group. For instance, OpenAI’s voice mode has become a rapidly adopted feature of the ChatGPT mobile app.

There are many workflows that can benefit from the technology, Jackson noted. “Workers who are most often in front of a computer may still prefer text interactions with AI, but many workers are on the go and their hands are busy,” said Jackson.

For instance, a construction site inspector could use a headset and smartphone to connect to an AI agent, and make an oral report while walking a site. They could ask the agent to share information on standards, and generate a report that’s logged directly into the system of record. Desk workers, for their part, may get tired of staring at a screen, and at times switch modes to use a voice AI assistant.

Jackson noted that AI voice is now closing the “uncanny valley gap” where humans are initially put off when confronted with a technology that purports to be, but is not quite, lifelike. AI voices are sounding more natural and fluid, and human voices can be emulated with just a few seconds of recorded audio. Realistic, unique AI voices that don’t imitate real people are commonplace and easy to create, and platforms that allow users to create content with AI voices provide a “nuanced set of controls to tweak emphasis, tone, warmth, fluidity, and other factors” that can create different vocal effects.

“It’s proven that it’s possible to listen to speech, process a response, and generate it with a realistic voice in almost real-time,” he said. “Now, vendors like Leena AI are looking to bring that experience to the workplace.”

More on AI in the workplace:

Why ChatGPT is crushing Microsoft Copilot

The car as Office: Mercedes integrates Teams, Intune and Copilot

Wait a minute — developers who use genAI tools are slower?

How AI is reshaping Slack

Hands on: An early look at Comet — and how AI browsers could change the internet

>