Conference

Jun 13, 2023

AI Summit London 2023

We had a great time at AI Summit London 2023. Check out our key takeaways from the event and the talks we were lucky enough to attend.

Jamie Dougherty

Machine Learning Engineer

Speechmatics Team

We had a great time at AI Summit London 2023. Check out our key takeaways from the event and the talks we were lucky enough to attend.

Our Highlights


Enabling Acceleration of the Digital Journey with Generative AI

James Smith & Wayne Buckley, Sprint Reply

Generative AI has captured the attention of most people around the world. We sat down at James Smith and Wayne Buckley’s keynote session to hear his thoughts on the topic.  

Focussing on business goals to help your employees understand AI and how to adopt it at scale, Wayne explained how the employee journey can be accelerated by applying AI. Technology enabled innovation is profoundly reshaping our world, for example robotics are driving unprecedented advancements over all enabling more informed decisions. Therefore, this change must be embraced and organizations need to be more adaptable to this change. It was discussed that while many want to learn more, there are very few that are ready to adopt AI at scale. The quicker the strategy can be executed, the better. Spending too much time on defining the strategy than executing it can put you behind your customer needs as well as your competitors.  

While in some cases it can be easier said than done, Wayne Buckley went into detail on Reply’s approach to how to adapt to AI.  

  • Prove the opportunity to your organisation: Some individuals will be less confident about it, so you need to demonstrate how it works, the opportunity, and what it would impact differently

  • Develop an innovation strategy: It is inevitable that the level of disruption will increase. It is imperative to manage this safely while at a good speed and always keep culture in mind. Keep the right aspirations and resources, so individuals know how to apply and lead AI solutions.  

Key Reasons to Prove the AI Concept:  

  • Feasibility: Prove AI can solve the problem; demonstrate AI is the right solution and you can achieve the impact you want  

  • Impact: Understanding the potential benefits and risks will help prove a more well-rounded case 

  • Learning: Understand AI technologies, technical, legal, ethical considerations  

  • Confidence: Build confidence for the organisation that AI can be a solution 

Case Study: EasyJet holidays and applying Generative AI  

EasyJet Holidays want to make travel easy, improve efficiency and create engagement and excitement. They want to find numerous Proof of Concepts to help them, and no one size fits all. This led them to use Chat GPT as a Chatbot to harvest LLM (Large Language Models) and use pre-existing LLM to reach a wider audience such as German and French. They have a cast digital library of information assets and how to make them accessible. This includes FAQ’s, policies, and them using Microsoft power AI to answer questions from the indexed website content.  

AI is the more efficient option when we discuss this next solution. On average, it took 22 minutes for an agent to process a complaint and find a solution for a customer, while an AI contact center can complete the same task within 30 seconds by email. Even with humans in the loop it is less than 5 minutes. Leading to having Avatars to check in when staff are unavailable to answer questions, building trust and a strong customer experience.   

More information on James Smith and Wayne Buckley and the talk here.   


The Bots and the Brave

Lord Holmes of Richmond MBE

Benefiting humanity or end to humanity? 

When AI is mentioned it’s common to ask the question, is AI benefiting humanity or is it an end to humanity? When considering this, it’s important to think about what governments and regulators need to do. 

As mentioned in the talk inclusion is one of the most important elements of how we adapt AI into the future. Essentially it is our humanities responsibility to ensure the AI era doesn’t go wrong. As Lord explains ‘’if we stuff this up, this won't be a failure of AI's, it will be a failure of us.’’  

More about Lord Holmes here.  

Conway’s Law and Machine Learning 

How generalists Thrive When the Rules Keep Changing  

Our very own Machine Learning Engineer, Jamie Dougherty related Conway’s Law and Machine Learning during a talk on the first day of the AI Summit. Mentioning ‘’we are often navigating the sea of potential pitfalls and dead ends of software development. In this world, it’s the generalists who can have the largest impact.’’ Using Speechmatics Ursa launch to showcase high accuracy transcription to create new life changing technologies. Comparing the Word Error Rate (WER) to other companies to showcase that having a birds eye view reveals new opportunities.  

Thinking about Conway’s law should not just be about organizations but also on expanding your own knowledge base to find new solutions.


And now for that sneak peek we promised...

Showing off Real-Time Translation  

We’re excited to have the opportunity to showcase our Real-Time Translation in...real time with all those who attended our booth. It’s one thing to hear about accuracy and speed but another to experience it. We loved meeting all the attendees, if you couldn’t make it to our booth you can still try our portal for free.  

Attending the AI Summit London was an incredible experience, and we are looking forward to AI Summit in New York City this December where we can meet more AI enthusiasts and have more amazing conversations on how our speech-to-text technology is making a difference in accuracy and speed everyday globally.  

See you soon!

In Case You Can't Make It