With nearly a year gone by since ChatGPT was made public, now's a good time to see which companies have successfully jumped on the AI wave to deliver value for their users.
As Vasi Philomin, VP of generative AI at AWS told TechCrunch in a recent interview: “We think every application out there can be reimagined with generative AI.”
Let's take a look at how innovations like vector databases, semantic search and function calls are enabling LLMs (Large Language Models) to go beyond chat.
Despite only entering the podcast game in 2018,
Semantic search has been made possible by new algorithms and databases that can capture and store the meaning between words and phrases. It lets users search intuitively, in the same way we might ask a real person where to find something – with descriptive phrases rather than by keywords. It’s super powerful because we tend to have an idea of what we’re looking for, but rarely do we know precisely which terms appear in what we want.
With this technology, Spotify's podcast user experience is only going to refine and improve. It's also a great example of how to use new generative technology to work towards an existing strategy.
Whilst Spotify took years to refine their search technology, for our skincare client Skin Rocks, we achieved a similar effect quickly and inexpensively with third-party technology: specifically the Algolia search API.
Algolia’s combination of semantic and keyword search allows users to search by symptoms they want to remedy or the results they’re looking to achieve – rather than just by product name. The app then returns users with the exact products that have been rated by Skin Rocks in-house experts. This feature is core to the Skin Rocks offering of helping people find products they can trust. And with 200,000+ active users 1 year after launch, it's clear that customers are getting the value they want.
The potential of ChatGPT for your business goes beyond questions and answers. “Function calls” allow your LLM-powered chat to interact with your internal APIs, meaning chatbots can suggest and complete requests for users on their behalf – all from the chat window.
Shopify, the e-commerce platform, is developing a suite of AI-enabled features under its sub-brand “Shopify Magic”. One of these tools, Sidekick (a custom chatbot), is trained to “know all of Shopify” and access the context, data, and expertise to generate personalised support relevant to your account. You can ask it to generate store-wide coupons, get insights and reports generated on your best-performing products all in a fraction of the time it would take you to navigate these functions and their documentation manually.
Revolut, the global neobank and fintech, partnered with AI company Aylien to empower customers with real-time news to increase engagement with their investment offering.
The Aylien News API provides real-time access to enriched, tagged, and structured news feeds. This means Revolut can provide personalised news for their customers’ portfolios and help them make informed investment decisions. As a result, they’ve seen an increase of 75% month over month (MoM) in users interacting with the global market news feature.
OpenTable, the online booking app for fine dining, was stuck with an unscalable content moderation strategy. Only being able to manually review 300 images daily meant that inappropriate content was being uploaded as part of user-generated reviews and not removed speedily.
OpenTable partnered with Clarifai to train custom AI models that ensure effective content moderation. This increased the number of images moderated from 300 to 5,000 per day (per moderator) and mitigated any legal concerns over unwanted content and copyright infringements.
Be My Eyes is an amazing innovator in accessibility. Founded in 2015, they connect blind or low-vision users to sighted volunteers and professionals through real-time, video-chat assistance. This lets those with visual impairment get assistance with daily tasks at a moment's notice, like identifying a product in the shops or navigating an airport.
Be My Eyes found that some users actually felt uncomfortable with the human aspect. In response, they’ve launched a Virtual Volunteer tool, powered by GPT-4 and capable of generating the same level of context and understanding as a human volunteer with image-to-text and voice. Users can then re-prompt the tool to get more information that is usable.
"Up until this point, we just hadn't seen a technological tool that would solve these needs quickly enough and accurately enough to launch something like the Virtual Volunteer tool"
At Planes, we’re big fans of understanding user behaviour. Amplitude helps get insights on your users at scale with product analytics and event tracking. Anyone who has used a platform like Amplitude will tell you that whilst they’re immensely powerful, getting started can be super overwhelming.
Enter the Amplitude Smart Assistant. With this AI-enabled feature users can now “Ask Amplitude” questions, in natural language, about their data and get back the chart with the answer they need. If that's not enough, the “explain” feature gives further insights to help understand complex data, dramatically reducing the barriers to getting started and identifying opportunities for optimising your customer journeys.
Ramp are using LLMs to go after the finance back office market. Ramp rely heavily on accurately interpreting the unstructured data of contracts, invoices and receipts. With LLMs, they can turn these inputs into structured metadata which they can then search over and summarise to drive insights for businesses about card usage and spending patterns to help cost saving. Pretty neat.
Koios helps job applicants and Talent and HR teams. With their proprietary AI, they cut out the cumbersome task of completing psychometric tests. All candidates have to do is provide an audio clip and Koios analysis gives the same insights to interviewers on cognitive ability and personality.
We worked with Koios to develop their product proposition and leveraged OpenAI's GPT when designing their reports feature. With GPT we could easily automate how we summarise insights and create personalised playbacks custom to each job listing. Automated insights + automated reports = winning.
If customer service headaches are your stumbling block then look to Intercom, one of the most successful chatbot companies out there. With their new innovation “Fin”, they’ve bet the company on the fact that AI is here to stay.
Intercom’s Fin enables companies to effortlessly integrate an LLM-powered chatbot. With rapid setup (little more than access to your existing help desk documentation) it can automatically resolve up to 50% of support questions, freeing up customer service teams to focus on more strategic plans.
Whether you or your company plan to build or buy with generative AI (or do a combination of both) we believe there’s a significant opportunity to innovate in this space. Generative AI can help make your technology more intuitive and open up new use cases for your products and services.
If you’re interested in exploring how AI could unlock new ideas on your roadmap then get in touch. We’re running free half-day workshops to answer all your questions and collaborate on what opportunities exist. We’re experienced with building, buying and everything in between.