Microsoft has introduced its flagship AI, Copilot, to the Russian messaging app Telegram in what might become a red-letter day for the AI race.
What does it actually mean for the present, and what could it portend for the future?
Okay, So What’s Actually Happened?
Microsoft has introduced an official bot version of its Copilot AI product to Telegram, which empowers users to search, ask queries, and chat.
The Copilot bot currently works in Telegram in the US and UK. It is available for free on mobile and desktop but requires a user’s phone number to operate, which it shares with Microsoft, apparently to validate that users aren’t trying to access it from the EU, which has stricter data regulations than the US and UK.
As reported by Windows Latest, Microsoft aims to expand “copilot-for-social”, that is, to introduce its Gen AI to social media apps.
Microsoft published a blog post providing details on the service:
Copilot is a personal generative AI assistant powered by GPT model and Bing Search, available within Telegram. Copilot for Telegram uses natural language to provide responses on a range of topics, from seeking knowledge to travel tips to sports updates to movie recommendations. Let Copilot enhance your Telegram experience with its quick-witted assistance and endless curiosity.”
Copilot for Telegram is similar to other Copilots but has some limitations. It’s restricted to text requests and cannot generate images. However, it can search for information on the internet. There’s a daily 30-turn limit, meaning users and Copilot can only have 30 back-and-forth daily exchanges.
Why Is This Especially Noteworthy?
Several AI companies now offer access to their LLMs through messaging apps. Google allows users to interact with its Gemini LLM via Google Messages on Android phones. At the same time, Meta has introduced Meta AI to its chat apps, including Messenger, WhatsApp, and Instagram messaging.
Likewise, in recent months, Microsoft has incorporated Copilot into several of its products and services. In addition to Copilot for Microsoft 365, there are now PCs with built-in Copilots, Copilots for business apps, and a subscription-based iteration that grants access to the latest AI models for $20 a month.
However, a tech giant bringing an LLM to a platform it doesn’t own is eyebrow-raising—at least, if that company isn’t named OpenAI and its service isn’t named ChatGPT. Similarly, Microsoft expanding Copilot beyond its own stable is novel.
This move could pioneer Copilot’s introduction to other social media platforms and messaging apps, as telegraphed by Microsoft’s “Copilot for social” project. While there are inevitably trust and compliance issues with having a separate organisation’s AI access a different company’s platform, it’s likely that Microsoft’s integration plans extend to the likes of Meta’s WhatsApp and Messenger, provided Meta is open to collaboration.
Going even further, could it also foreshadow its arrival on competitor UC and collaboration platforms?
Could Copilot Arrive On Other Major UC&C Platforms?
The impulsive answer might be, “No, that’s ridiculous; why would Microsoft’s rivals shoot themselves in the foot?”
The idea of a Copilot bot on Cisco Webex or Zoom’s new Workplace platform doesn’t make sense at first glance, especially with the former’s AI Assistant and the latter’s AI Companion evolving into strategic focal points for the vendors. They also already perform much of the same functionality that Copilot does in the Microsoft ecosystem, from transcribing and translating meetings and producing action items to drafting messages for chats and emails.
As the AI race continues to heat up, any slight advantage or distinguishing selling point these businesses can illustrate to customers could be invaluable. Zoom’s AI Companion doesn’t require extra cost within Workplace, for example, while Copilot 365 does.
However, two trends suggest the notion of Copilot bots landing on Zoom (or AI Companion bots on Teams) isn’t as outrageous as it might initially seem. The first is the dynamism of third-party development in adapting and refining AIs to suit their own individual or organisational needs.
While there isn’t a native Copilot application for Zoom or Google platforms, developers may leverage Microsoft’s Azure AI Studio to create custom integrations that could bring Copilot-like functionalities to these platforms. This means that while Copilot itself may not currently be directly available, its underlying technology could be adapted for use on other platforms by third-party developers.
However, the more compelling 2024 trend underpinning the possibility of one company’s AI being open to integration with other platforms is the growing focus on interoperability between UC platforms around meetings and messaging.
The idea that workers could directly join Zoom or Google Meet calls from their Microsoft Teams app might have sounded similarly ridiculous to AI interoperability a few years ago. However, last July, Microsoft introduced the ability to join meetings set up by third-party services, including Zoom and Google Meet, directly through their Teams calendar.
In January, Cisco, Microsoft, and Samsung jointly announced a set of meeting room solutions. The combined room kits entail Cisco devices powered by its RoomOS software integrated with Samsung smart signage displays featuring Microsoft Teams with Front Row.
Meta has adapted WhatsApp to become interoperable with third-party messaging services while preserving its end-to-end encrypted messaging. While this move was facilitated by the European Union’s Digital Markets Act, which requires digital gatekeepers’ messaging services to be interoperable with third parties, that in itself could also foreshadow how future AI legislation, currently in its most nascent form, could mandate that different LLMs be interoperable with other platforms.
However, customer demand may be a more pressing factor behind interoperability than potential regulation.
Customer Demand For Access To Their AI, Anywhere, Anytime
As Microsoft irons out Copilot’s kinks as a productivity assistant and more users adjust to leveraging Copilot in their everyday work lives, whether through its enterprise-based 365 version or even through casual use of the consumer Windows or Edge iterations, utilising the service could gradually become a new normal. With a new normal comes a new weight of customer expectations.
A similar dynamic could well emerge with Google and Gemini’s integration across the Chrome browser, Google search engine and Google’s enterprise-targeting products like Workspace and Meet.
As with Copilot, Gemini’s user experience on Google is deeply flawed in its current state, but technical updates, enhancements to product sophistication and growing user adaptation could see Google’s Gemini and Microsoft’s Copilot become as natural and integral to people’s work lives (and personal lives) as Wi-Fi and smartphones. Its transformative potential is undoubtedly there.
When access to our Wi-Fi and smartphones abruptly disappears, our dependency on them now is such that it can be immensely disruptive. If your AI service of choice suddenly vanished—which, in the future, might have years of data about your working style and preferences (simultaneously an appealing and terrifying thought) with tailored recommendations and insights to support you—it might end up being similarly jarring.
Users might grow frustrated that the AI-powered meeting assistant they use in internal Teams calls isn’t available when they join external meetings via Webex or Zoom, or vice versa. They might be exasperated that notes they’d prepared in planning calls within their organisation aren’t readily available or updated when joining the external call, for example, or wish to refer to an in-house report that would have usually been available for the AI to produce at will in an internal meeting.
These are problems that a Copilot bot for Zoom or a Zoom AI Companion bot for Teams could potentially address. Imagine bringing your Zoom AI Companion, Copilot or Gemini assistant with you to external meetings, or virtual events or conferences, or on a personal messaging platform. Inevitably, this will raise security and compliance issues, but the scope for innovation is there.
As one industry colleague UC Today spoke to at ISE Barcelona this year said, there is a “grey area between competitor and collaborator” in UC.
As outlandish as it might seem midway through 2024, if customer demand for being able to take their AI assistant of choice with them across platforms and apps amplifies and eventually outweighs the benefits of exclusivity, there might come the point where it’s in vendors’ best interests to metamorphose from competitors into collaborators and oblige them.
