2021 has been a big year for us. Against the challenging backdrop of the global pandemic, Re:infer and its people continued to thrive - taking on new opportunities, new customers, changing and growing as a business.
As we look forward to what 2022 has in store, we wanted to get together to remember some of the key milestones we’ve hit over the last year, and tease a few of the things we have planned…
Re:infer secures $7.6m in funding
This year we were very excited to announce that Re:infer has raised $7.6 million from investors, including IP Group, Crane Venture Partners and Seedcamp.
We’re already putting this investment to good use. We’ve expanded our research, engineering and commercial teams, and continue to do so with many new roles and opportunities. As ever, we are also exploring new applications for Conversational Data Intelligence, finding new ways to create fresh value for our customers.
Another huge thank you to our investors, friends and partners - we couldn’t have gotten this far without you!
A Gartner Cool Vendor in Conversational AI and NLT
We were delighted to have been recognised as a Gartner Cool Vendor this year. In the Cool Vendors in Conversational and NLT report, Gartner analysts counted Re:infer among a select group of innovative solutions providers tackling challenges in reinforcement learning, decision intelligence, natural language understanding, and conversational AI.
We see Gartner’s recognition as an important step in our work to build the category of Conversational Data Intelligence - helping enterprises create business value from their business conversations. In the new year, we’ll be building on this recognition, growing our ecosystem of partners and customers, as well as increasing awareness of the value hidden in business conversations.
Re:infer outperforms Google AutoML
When we speak to businesses about their struggles with the massive amounts of communications they send and receive, they often tell us how they’ve tried to solve the problem themselves. We’ve heard from countless transformation leaders about the challenges of building their own custom solution, combining natural language processing and machine learning to try and understand and improve their business conversations.
More often than not, such projects struggle to reach the finish line. But we were curious to see what the output would be if a custom solution successfully made it into production. That’s why we decided to go head-to-head with Google AutoML - one of the strongest and most popular cloud NLP solutions out there - comparing our performance with theirs on a comprehensive communications dataset.
Our very own Machine Learning Research Engineer Arthur Wilcke performed the test, and you can read the full results on our Developer Blog. But just to summarise, we were delighted to see that the Re:Infer Conversational Data Intelligence platform outperformed Google AutoML on average by 5-10% on every benchmark.
We put a lot of effort into ensuring our platform is fast to train and deploy, so we were also extremely proud to see that model training on Re:Infer is approximately 200x faster than on AutoML!
Re:infer’s performance has only improved since this test, but we feel it has really vindicated our approach and shown that a specialised Conversational Data Intelligence platform is the best way to solve the challenges of enterprise communications.
It’s been a great year for starting work with new enterprises and strengthening our relationships with existing ones. We’re delighted to say that we’re ending 2021 with many great new customers - including another one of the world’s largest international banks. We have also extended and deepened our involvement with long-standing customers such as Hiscox and Farfetch.
It’s very gratifying to see that, once an organisation has experienced the value of Conversational Data Intelligence first-hand, they want to embed us even deeper into their business. We’re constantly exploring new use cases and applications with them to see how much value we can help them extract from their conversations.
The Re:infer partner ecosystem
Our network of partners and alliances is key to our success, ensuring our customers have access to all the tools and expertise they need for scaling digital transformation.
We learnt a lot from our partners this year, and we’re delighted to have signed new agreements with seven leading consultants and integrators - including Reveal Group, Future Workforce, Automation Anywhere and ISG.
It’s great to see a strong and thriving ecosystem of providers starting to coalesce around Conversational Data Intelligence. Organisations are waking up to the value of conversations and competitive efficiency in the enterprise. We’re now strongly placed to deliver on that promise.
The Re:infer team grows
One of Re:Infer’s core value propositions is helping customers scale up and unlock the value of their automation and digital transformation initiatives. But as we take on more customers and partnerships, that also means we’re constantly scaling up as a business ourselves.
In 2021, we grew into a company of more than 25 AI experts and engineers, crack developers, customer success, sales and marketing wizards.
We were delighted to welcome back Harshil Shah to our team of AI researchers following the completion of his Machine Learning PHD at UCL. Harshil has been very busy keeping us informed of the latest developments in machine learning and NLP, which is crucial for maintaining the top performance of the Re:infer platform.
We were also very happy to welcome Florin Buga, who has joined the team as a Senior Frontend Engineer. Florin has been hard at work driving and building some of the great new features that have arrived on the platform this year - most recently delivering our very well-received Dynamic Thread View functionality.
On the Commercial side, it’s been great to have Phil Hughes join us as our Head of Discovery. An experienced RPA and automation consultant, Phil has been playing a vital role working with customers to understand their needs and requirements, and helping them discover all the ways they can deliver more value with Conversational Data Intelligence.
Our Sales and Marketing team has also been scaled up with the addition of Tom Golebiewski and Aaron Crossey, who have been doing a great job leading prospecting efforts, and driving our content marketing activities respectively.
We have no intention of stopping there. We’re always looking out for fresh talent and experienced veterans who want to work at one the UK’s most innovative AI companies and unlock the power of AI for everyone.
Nothing has changed more over the last year than our Conversational Data Intelligence platform. Our development team has worked hard this year, and in collaboration with our customers, to improve Re:infer and greatly extend its capabilities. Here are some of the highlights:
Re:Infer machine learning models are now able to read messages and datasets containing multiple European languages. Users can now train models on data that has a mix of European languages, even within the same conversation.
In 2021, we started by adding English, French, German, Spanish, Italian, Portuguese and Dutch, but are working to expand this list even further.
We’re very proud of our Model Rating feature, which has made training models in Re:infer faster and easier than ever. Model Rating automatically analyses models as they’re being built and rates them based on their accuracy, balance and performance. It then reports back a holistic score that gives SMEs an accurate picture of the model’s health, as well as the actions needed to improve it.
We’ve augmented the platform’s analytical capabilities to allow users to apply multiple label filters at once, exclude certain labels with a filter, search with label filters, and more. This feature was in high demand among our customers, and now enables them to perform deeper analysis of their communications data.
We’ve transformed our approach to entity extraction, building a whole new entity model on a transformer-based architecture - a type of deep learning that learns to only pay attention to the most relevant parts of an input.
Since rolling out the new model, users have been enjoying greater accuracy, faster training and better validation statistics, helping models enter production and start generating value much quicker.
Our development team is hard at work on our next big update. We’ll soon be adding a new Quality of Service feature to the platform, giving customers a whole new metric to measure the quality of their customer interactions - with every message contributing to an overall Quality of Service score.
Our commitment to hybrid working
Safety is our number one priority. Over the last year we’ve continued to adapt our working arrangements in response to the pandemic and the needs of our people. We know how difficult this time has been for everyone. That’s why we want to offer the most flexible working conditions possible so everyone can work safely, happily and productively.
Re:infer is fully set up for hybrid working - whether you want to come into the office every day or work remotely from home. We use collaboration tools like Slack, Zoom, Trello and Notion to make sure we can all work together virtually, in-person or both.
Despite the challenges, this hybrid system has worked remarkably well for us, and we’re fully committed to seeing it continue. But that doesn’t mean we’re neglecting our office culture. Over the last year, the team has done a great job of keeping our tradition of Friday lunches alive. We’ve also commandeered the basement on occasion for our regular board game evenings. We’re delighted every time a new or familiar face drops in to say hello.