The Aveni UX Designer explains how they use machine learning to help financial advisers capture client conversations.
Pitch Aveni to us in 100 words or fewer. Go!
Aveni powers automation and insight directly from the voice of customers.
We use machine learning and AI to analyse conversational data and pull out key risks and information. We don’t just help people understand their clients and agents better but we use speech analytics to provide actionable outcomes and automate processes for greater efficiency. Allowing advisers and businesses to retrace their steps, generate more accurate reports and track progress.
Financial services are changing and manual processes are being left behind - Aveni helps advisers and businesses maintain focus on the customer, whilst the details fall into place.
Talk me through your experience of working at Aveni and what you’ve learned from it so far. What was your biggest challenge?
My biggest challenge working with Aveni was gaining that start up mentality! We have a great team and everything has moved and grown so fast. I’m so proud of what we have created so far but in an agile start up there’s a lot of trial, error and lessons. Remembering that imperfection is a part of improvement helps me shoo off that impostor syndrome and focus on how far we have come.
Would you do anything differently next time?
I would have taken advantage of our time in an accelerator and done lots of fun guerrilla testing, it was a lively office and I would have liked to have had the confidence to swan up to the different teams with some prototypes and ask for their thoughts. It’s great having all these testing sites but there’s nothing like huddling around a table and getting that feedback. Also, I’ve found more time recently to document my process and I wish I had put more effort into annotating my work earlier. It’s been helping me improve my communication skills and justify my ideas.
What has it been like to work with IFAs? How can you help them train the models you use?
Working with IFAs has been fantastic, I think the industry is ready for a change in process and everyone we have worked with has been so enthusiastic and helpful. We are an Agile team so we like to maintain transparency and keep our users in the loop, and we have a unique opportunity to de-mystify ML with a simple and usable interface. Giving users the opportunity to learn by using the software is important to us when getting people on board.
Briefly describe your ‘toolkit’ – what techniques and approaches work for you when combining UX design with Machine Learning (ML)?
I wouldn’t say I have an ML-UX toolkit, really. I’m lucky to work in a close-knit team so I can sense-check ideas with our ML experts anytime. I think open communication is so important when developing these complex platforms. Sometimes ML concepts are so tricky for non-experts but sitting down together and looking at whether a UI solution fits with the data helps make good progress quickly.
Did you find any useful Machine Learning books, blogs or cool stuff you would recommend for other designers?
I would recommend reading a lot of (recent) Medium articles about data visualisation and looking out for JS and react data vis libraries for inspiration, usually they have CodePen examples you can play around with to get ideas flowing. Other than that I would encourage other designers to join discord groups like Designbuddies. These groups have loads of great mentors and experienced designers that are keen to give feedback and advice quickly. A total game changer for me, especially since we have all been cooped up at home!
(Mhairi also recommends the Machine Learning and design website.
In your opinion, what's the biggest impact ML will have on UX design and research?
I honestly have no clue, but I’m excited to see what happens. I would love to take part in a hackathon where we make a tool to generate user profiles from interview audio or something!
I do hope ML can be used to improve accessibility testing and standards using image understanding etc, that would be really cool.