The Enterprise LLM Is Here, And It’s Going To Change Everything
I’ve had a lot of cool and interesting experiences in my career. One of them was a brief stint working alongside Jeff Dachis, the co-founder of the legendary digital firm, Razorfish and currently founder and CEO at One Drop. Back then, we were sizing up how social media was transforming the world again (this was 2007). He recalled the last time he had been at the forefront of such significant digital transformation was at Razorfish. It was there that a common mantra was coined:
Everything that can be digital will be.
The genesis of that phrase originates as far back as the late 90s. So here we are again, with AI poised to change the game and how we view things like content creation while forever changing what we had previously thought about things like Chatbots.
ChatGPT, has introduced potential Large Language Models to the public faster than nearly any technological revolution that came before it. And like many previous technological revolutions that start in the hands of consumers — it inevitably bleeds into the enterprise. Platforms such as Slack, for example, borrowed generously from consumer mobile and social apps, which consumers had become accustomed to — and now they are platforms that exist within all kinds of business ecosystems and have become for work what oxygen is like for living. Brian Solis, head of global innovation for Service Now, describes AI initiatives at the enterprise level as a way to supercharge workflows:
it’s time to start thinking about genAI as an enabler for augmentation and exponential performance. Strategies must run in parallel for automation + augmentation. One way to do that is to use AI-to supercharge workflows across the enterprise. Additionally, companies should explore areas where private, smaller, secure language models to empower employees with generative AI capabilities in customer service, forecasting, marketing, predictive models, and so on.
Enter Enterprise LLMs
We’re in the early stages of organizations feeding their data into large language models that serve conversational experiences which pull directly from the organization’s private and secure data alone or combine it with data scraped from the open Web. Global consulting behemoth McKinsey recently announced the launch of their Enterprise LLM application, “Lilli”:
“If you could ask the totality of McKinsey’s knowledge a question, and [an AI] could answer back, what would that do for the company? That’s exactly what Lilli is,” McKinsey senior partner Erik Roth, who led the product’s development, said in a video interview with VentureBeat.
Lilli represents what is going to become the norm for many organizations and enterprises over the next five years or so, and that is taking their vast data sets, training that data into a large language model (privately and securely), and adding an experience layer so employees, customers, patients, etc. can interact with it. In short, Enterprise LLMs have the potential to take the original digital mantra and LLM it.
Everything that can be LLM-ed Will Be
Another example comes from Morningstar. (Disclaimer, the example below includes technology created by the company I work for — Soul Machines). Morningstar CEO Kunal Kapoor is demonstrating an LLM Morningstar created while using Soul Machine’s digital avatar technology as the experience layer. In this case, the Enterprise LLM is called “Mo.”
Jeremiah Owyang, founder, investor, and Silicon Valley insider describes the opportunities Enterprise LLMs present to organizations as a way to embody the organization’s ethos:
“Companies are launching enterprise LLMs to extract employee knowledge, experiences, and insights. By gathering data from various sources, they create an AI that embodies company ethos, accessible via simple chat, constantly evolving.”
He goes on to lay out the benefits of “every company having its own AI”:
“Once the LLM ingests the documents for internal usage, it will be tested with employees in a safe environment. Then it will be segmented for marketing, sales, product interface, customer care, and beyond.” “This enterprise LLM paves the way for every company to have its own AI, reducing reliance on tech giants that offer it as a service.”
It’s not just McKinsey leaning into a proprietary Enterprise LLM — PwC is also evolving its Enterprise LLM with more generative AI capabilities fusing the benefits of Open AI’s GPT engine combined with proprietary PwC data — the result? “Chat PwC.” Scott Liken, Global AI and Innovation technology leader at PwC, describes the effort as such:
“We made the initial wave of investment in ChatPwC, which is our interface to leveraging the large language model from OpenAI and then others as needed,” said Likens. “To do that, we had to make it valuable to the work we do every day, so that means bringing in some of our own [intellectual property].”
It’s no accident that large consultancies such as McKinsey and PwC are amongst the first to evolve their company AI into more sophisticated and modern Enterprise LLM-based experiences — firms like this will be instrumental in guiding their impressive client base through similar transformations. They can point to the knowledge they have gained by doing it for themselves.
Generative Experiences: Combining Scale With A Human Touch
Enterprise LLMs partially illustrated in the above examples will allow large organizations to engage key stakeholders in a natural, conversational way — and do so at scale — securely and with enterprise-grade infrastructure powering all of the operations behind the scene. But none of this will be visible to the person at the end of the experience. The magic that comes with the experience will be the feeling that you’re having a natural conversation or interaction with something that you know is not human — but feels human-like.
Generative experiences such as these can transform the paradigm for how we interface with systems. For the past thirty-plus years — we’ve been accustomed to interacting with “GUI” (Graphic User Interface) regarding human-computer interactions. But now, GUI will be making room for “CUI” (Conversational User Interface). For proof of this, see the 15M round of funding Voiceflow recently recieved — positioning itself as “the Figma of Conversational AI”. With the consumerization of the LLM, taking chat and Conversational AI to the next level and combining that with personalized data, the customer/employee experience will evolve rapidly. And for organizations? Remember this simple phrase:
Everything that can be LLM-ed Will Be