CorpVerse Series #7. Corporate Metaverse. AI Assistants
As a renaissance of chatbots, we note not only the phenomenal ChatGTP success but also a significant shift in the interaction paradigm from scripted flow to exploration. Text as an interface is alive
In the IT industry, we have been trying to embrace the usage of AI assistants and chatbots for the last ten years. In these attempts, we have come a long way from the hype with a lot of frameworks, design tools, and startups to a tenfold drop in expectations from the real capabilities of the technology. Most of the "real" bots have been reduced to simple scripts that work on knowledge bases and relatively simple logical chains with intent recognition. The request to "call for a human" has become the norm in "smart assistants" dialogs. Microsoft shut down Cortana, and Amazon's most successful digital assistant, Alexa, was probably a colossal failure.
In that context as a renaissance of chatbots, we note not only the phenomenal success of transformer engines like ChatGTP and Anthropic but also a significant shift in the interaction paradigm from scripted flow to exploration. The generative nature of the new technology and the ability to integrate it into any text field was successfully sold to the mass user. Its answers are interesting, but certainly not necessarily accurate - and that's OK! It has an "open script" that developers can enhance on the backend with fact-checking and contextual data. And it's great! Text as an interface is alive again.
🔮 Insight. Revive the stillborn
Communication vendors and startups provide "standard" Metaverse bots (e.g., meeting secretaries with built-in functions and minor customization options).
Despite being promised by the Metaverse evangelists, the worlds attributed today to the Metaverse are rather empty than full of life. But this is not surprising, because even outside the Metaverse we do not have many digital examples when a sufficient number of simultaneous users could gather in one place and at one time, while actively exchanging interactions with other participants.
The Metaverse expectations paradox. We have social networks with hundreds of millions and even billions of users, but we do not expect all of them or a significant part of them to discuss something effectively on the same page at the same time. For many years we have had media sites with content watched by tens or hundreds of thousands of viewers at the same time, yet we have just begun experimenting with UX for real-time interactions between audience and content producers. We have online games with huge numbers of simultaneous players, but we still don't know how to bring them all into one game session and virtual space. However, all these issues unresolved in isolation we expect to be resolved in the Metaverse.
We are confident that this "paradox" will be resolved, and not in the favor of human dreams. In reality, we have a very, very small number of scenarios requiring bringing any significant number of people into one audio-visual space. In truth, we have more scenarios in which we almost don’t need anyone―and it’s enough just to create the illusion of life around (for example, make the top executives feel like Freddie Mercury by drawing a stadium of employees in front of them): “fill” the world with a simulation of life, place simple "characters" in it with minimal customization options and minimal intelligence, or autonomous shadows (ghosts) of other people, as has been done in car simulators for a long time.
In corporate metaverses, everything is both simpler and more complex. On the one hand, companies are usually smaller in the number of employees than social networks, and activities are divided into separate small departments, so ultimately, the task sounds like bringing one or two hundred people together in one mixed space (it is already possible). On the other hand, you know all these people personally, so you expect them to be real, and you will notice if half the team is missing. And if the new corporate metaverse is empty, then how is it better than the new messenger, in which there is also no one (at first)?
In this vein, we expect that communication vendors and startups will start offering standard "bots" that will fill in pauses, record a meeting, tell you about the weather, ask how you are, wish you a happy birthday, etc. Many of them will be based on ChatGPT and its analogs in the coming years, giving employees the illusion of having a sufficient number of available interlocutors. To make a shift towards intelligent assistants in the Metaverse, two steps remain 1) connect large linguistic models to corporate knowledge bases and automation tools, and 2) turn them into NPCs (non-playable characters) - avatars that fill virtual and mixed spaces with utilitarian functions. Your virtual office will have ready-made gardeners watering the flowers, but not knowing anything about the company, and receptionists connected to the calendar, probably not knowing anything about the flowers.
⚗️ Experiment. Welcome our new Teams member
HR departments build the first AI avatars with low/no code for new employees onboarding.
According to the classics, low/no-code tools for corporate environments are mainly designed for business analysts and other roles working with data. With the advent of large linguistic models, the coverage expands to “supporting” roles that are faced with repetitive requests from colleagues. This is the administrative contour of the company: lawyers, accountants, IT support, and HR employees, who, on the one hand, have already managed to accumulate knowledge bases with questions, answers, templates, and instructions, and on the other hand, they are full of futile prospects of automation with the previous generation of chatbots.
One day you come to work and find an email greeting a new employee. He or she has an avatar, name, email, calendar slots, and even a "before" history. Perhaps he is a new HR employee who accompanies newcomers and helps them understand processes, hierarchies, and paperwork. Or maybe she is a new help desk employee who can help you access internal systems and set up device policies. Or a lawyer who prefers “they” (and politely corrects you if you suddenly mix up the gender), who will help you customize the contract and carefully listens to all the wishes of your partners and clients.
You can write to all of them directly in Slack or Teams (or another tool deployed in your company), you can even call and talk or hang out with them in Horizon (!). Somewhere in the core of your consciousness, you are surprised at the diligence, speed of response, and politeness of employees, you will not even argue if they become employees of the month or year!
A terrible secret is being stubbornly hidden from you: each of these employees is a virtual bot with avatar representation programmed with a low or no-code platform by real human employees who unexpectedly became "zero-coders" and perceive each of the created bots as a new pair of hands. Finally, they can give the routine to machines, and focus on creativity themselves!
⚙️ Practice. The unlimited front row
All-hands meetings, courses for onboarding, and kick-off processes built as interactive Metaverse experience.
The persistent remote and hybrid working modes force corporations to abandon empty office spaces, and with them large halls for company-wide meetings. All-hands meetings are transferred to digital environments on streaming platforms with new interaction tools: picking up questions from the audience, real-time reactions using emojis, and virtual spectators arrangements.
The most advanced companies experiment with the Metaverse platforms for interactive meetings. Employees in a company like the Meta Platforms are faced with a simple choice: watch the old-fashioned video with Mark or put on a Quest helmet and get the opportunity to watch the founder's speech from the front row and even shake their courageous virtual hand and get a digital autograph. This is the third mention of NFT in our report, one more to go.
With the Metaverse, it becomes possible for each employee to get into the front row of the parterre at a large corporate meeting, everyone can take a virtual selfie with the legend's virtual avatar. And the price tag for inviting a virtual artist, with whom you can fly around virtual worlds together, will probably be an order of magnitude less than for a physical exclusive concert.
Perhaps one day everyone will be able to hear a personal answer to their personal question, but we will leave that for the next sections.
🚸 Risk. The smart ears and eyes
3rd-party bots added into corporate meetings pose security risks.
While you visit conferences and meetups and scribble revelations about your innovative attempts in Telegram chats, telling colleagues across the industry how neoteric and fervently you are now living with avatars, NPC characters, and zero-code assistants running errands in these Metaverses, your security guards are becoming suspicious and starting asking unpleasant questions.
For example, how do you know that these third-party "smart" bots are not digitally earjacking while you are holding a top-secret meeting in the Metaverse? After all, "Was zwei Menschen im Metaversum wissen, weiß auch das Meta-Schwein", and then everyone knows what the meta-pig knows. Woo-wee-ee. This seems to be your first serious conversation about security in the Metaverse. 🐖
You’ve seen enough on the Producthunt, read about ChatGPT, and, sorry, added another Albus bot to Slack. “Had you ever asked,” the security officer continues pressing you, “How many intermediaries are there reading your requests? What do they do with the data they collect? How is it processed? Do you already have bards instead of brains?”
📧 This is Human Spectrum Lab, a newsletter about Exploring the Human Spectrum and Shining through the Future.
In this series of posts, I share pieces from the Corporate Metaverse report (v2) written by Constantin Kichinsky and based on the original research on the topic done by Constantin Kichinsky, Aleksei Kalenchuk & Ekaterina Filatova in 2022.