Introducing Co:Driver, Peak’s new generative AI capabilityBy Ira Dubinksy and Rhys Davies on October 20, 2023
It feels like everywhere we turn there is another article or announcement about generative AI.
There’s a lot of excitement about this emerging technology and it’s not hard to see why: ChatGPT and other systems based on Large Language Models (LLMs) generate rapid responses that are as surprising as they are impressive in their ability to seem human. When it comes to imagery, tools like DALL-E and Midjourney have spawned countless cultural artefacts in the form of unusual combinations of artistic styles and subjects.
An example of image generated by DALL-E
But despite all of this hype, we’re really only scratching the surface of what generative AI can do — especially in business. To date, generative AI applications have overwhelmingly focused on the divergence of information: they create new content based on a set of instructions; the user enters something into a prompt and they get a response.
When we use generative AI as consumers, our objectives are about having fun, sharing something memorable or saving a bit of time. The applications that allow us to do that have quickly captured the imagination of millions of people around the world.
When it comes to the business application of generative AI, the focus must be about making things more efficient, driving growth or improving the quality of our work. There are some generative AI use cases for business that do just that. For example, sophisticated chatbots deployed in customer service, tools that assist in writing programming code or tools for content generation in marketing. But, the incremental value of generating new content from prompts will ultimately reach a plateau. Applications of generative AI that fundamentally enhance and adapt our business workflows will bring a much greater step change in productivity.
And that’s why we’re so excited to be unveiling Co:Driver, the new generative AI product from Peak.
Co:Driver combines a fine-tuned LLM with a user’s existing AI applications and business data. Co:Driver continuously searches for information, efficiencies, opportunities and anomalies of potential interest to the user.
These are surfaced as bite-sized contextual recommendations, with the user able to choose their preferred next step from a number of AI-generated actions. And through feedback from users, the recommended actions will improve over time.
As well as this, using outputs from a user’s existing AI applications on the Peak platform, users can ask questions about their business in natural language, such as how much stock there is in a certain location or which products are the best performers.
Just like the navigator of a rally car who sits in the front passenger seat, Co:Driver’s job is to navigate, tell you what’s ahead and what obstacles to look out for. It improves your performance, boosts confidence and brings knowledge and expertise to the job of driving a business forward.
A new paradigm for generative AI in business
Co:Driver is part of a new paradigm emerging when it comes to applying generative AI in business, and we think it will be transformational. The future of generative AI in the enterprise will see it facilitate an innovative and exciting new means to interact with your data and get value from it.
Generative AI will create a new layer between business users and the systems and data that run their operations, bringing together data from across many different formats and making sense of it autonomously. In doing this, the power of large language models will be used for data synthesis more than for content generation.
With this innovative approach, generative AI can provide business users with real-time, highly-relevant recommendations at their fingertips. Opportunities, challenges and anomalies related to stock holding, pricing or customer experience can be surfaced automatically and help leaders make timely, accurate, AI-driven decisions.
Here’s a practical example…
- Imagine a business that stocks and distributes construction materials such as lumber, mouldings and fasteners
- The business might have thousands of different products across a network of distribution centers and warehouses; they regularly fill orders from customers and re-order stock
- The company is using Peak’s Dynamic Inventory application to generate an AI-powered demand forecast and AI-powered recommendations on the correct quantity of stock to hold for each SKU, in each location
- Users typically review and action stock targets, at an individual SKU, category or sub-category level. Sometimes users manually scan the data for outliers, problems or opportunities
- Now imagine that, instead of manually reviewing the recommendations from the AI, the most important actions you need to take are surfaced automatically by Co:Driver. Exactly the information you need, when you need it
Your own AI
Up to now, discussions about different types of AI typically delineate between:
Predictive AI: uses machine learning and statistical algorithms to forecast trends, patterns and predictions from each user’s unique data set, and;
Generative AI: learns the patterns and structure of input data (e.g. text, images) and then generates new data that has similar characteristics.
Often these are presented side by side as if the two shall never meet. In reality, to get the most value for business they have to work together. Customers that work with Peak are building their own AI by training proven business applications on their data and business logic. With the addition of Co:Driver, the value of Peak’s applications is supercharged and they become even more powerful than they would be on their own.
Generative AI can be layered on top of predictive AI providing a new layer to interact with your data and your AI, one that is agnostic of the format of the data. This approach will become transformational in its ability to direct business teams on where to focus their efforts, thereby improving decision-making, boosting efficiency and delivering higher ROI. The combination of fine-tuning an LLM and Peak’s proprietary predictive AI applications will redefine the use of AI in business.
Co:Driver will be available in private preview in January 2024 — register your interest below.