DataSine CEO on the value of machine learning and AI in marketing
Everyone’s talking about artificial intelligence and how it can be used for commercial gain in the marketing world. One London firm on the cutting edge of the AI zeitgeist is marketing technology startup DataSine.
The company’s content personalisation platform, Pomegranate, is a collaborative AI-powered campaign platform that tailors content to personality.
Fresh from a £4m funding round to boost its expansion into new markets and sectors, we caught up with DataSine’s founder and CEO, Igor Volzhanin.
We started by asking him why machine learning was an important factor in the marketing business?
Marketing is an industry that has wholeheartedly embraced the big-data revolution. The data ecosystem within marketing is very mature and access to systems via API is very common.
This unlocks the value of machine learning and AI, because it requires a lot of training data to make these approaches worthwhile. Marketing tools which incorporate AI capabilities can make sense of the complex patterns in user engagement data to surface insights and empower marketers to build better campaigns.
Take A/B testing as an example. Marketers spend hours setting up competing campaigns in order to understand what content might do better.
AI can replace these efforts by providing instant analysis on how well a particular image, or button colour will do, leaving marketer with more time to do what truly matters, being more creative and focusing on how to better engage with the customers.
What is the killer win for companies using machine learning as part of their marketing strategies?
Machine learning helps companies create more personalised content and which ultimately builds more trust and engagement with their audience.
Personalised content has been proven to lead to increase in sales. We have helped companies increase call centre sales by over 71 per cent by personalising call centre scripts, powered by AI. For a large digital bank we increased sales via email personalisation by 59 per cent.
Which industry sectors are embracing the technology fastest?
During the first four years we focused on bringing our technology to financial institutions. We are now working with banks and insurance companies in four countries - France, Belgium, Russian and the UK. As success spreads, other industries are showing great interest, including telcos, the non-profit sector and e-commerce.
What are the key challenges they face when adopting new technologies like AI?
Arguably the biggest challenge facing marketers towards is the time it takes to learn all of the new tools coming to the market. With the marketing stack becoming more and more technologically driven, it is difficult to stay on top of the latest technology trends.
At DataSine, we decided to address this challenge by creating a product which is intuitive to use and which does not require any additional training. Although it is easy to use, it provides powerful insights that can be used instantly to improve content. Such is the power of applied AI in marketing.
How do new technologies like this help companies to understand customers better?
New technologies can use existing data to build new insights about customers. Our platform uses first party data, such as previous transactional and engagement data to provide new insights about customer preferences, including what content may resonate with them, what the best communication channel may be and how to best engage with them in the future.
The best thing about AI-powered tools is that they allow continuous learning. The more data you have about your customers, the more accurate personality profile you receive about each individual customer. This ensures continuous improvement of content and results in consumer loyalty and long-term satisfaction.
Are there any regulatory challenges to overcome, such as GDPR, etc?
We have always focused on first party data, by that I mean data that companies already collect in their regular course of business. Our customers do require to seek consent to use their data to generate additional insights, which is great since this provides an additional way for companies to educate their customers on use of data and to make sure both parties are aligned.
What is the biggest challenge you face when discussing using AI solutions with companies?
The biggest challenge is to show that it works to drive real-world outcomes. There are a lot of MarTech solutions that appear to provide value, but showing actual business impact is always difficult to determine.
To that end, we’ve had to run extensive statistically significant tests to show business impact and to write up case studies that could be shown to new customers.
How are you managing the ethical issues/concerns companies have in using AI?
As a company working with new technology like AI, we take pride in surfacing ethical concerns around usage and have encoded ethics as a core value of the company. There are two main areas of interest, one is the use of data and machine learning in a marketing setting, and the other is about algorithmic bias.
In the former, we see the main consumer concern to be that of consent, generally customers are concerned about whether their information is being shared to other people.
By helping companies analyse their own data, we see ourselves as helping them achieve strong first-party consent with their customers, without recourse to third-party data providers and ad networks.
With regards to algorithmic bias, we are strongly committed to opening black boxes and making our algorithms explainable. This has helped us reduce biases in the data, which are generated from our biased world, as we are very intentional in what features our models are allowed to use.
This helps prevent us from accidentally reinforcing these biases with our products.
What is keeping you up at night where AI growth is concerned?
I find that the most concerning cases are focused not on new developments in AI capability, but the application of very well studied AI capabilities, without properly assessing the potential impact of these systems.
The commoditisation of AI tools drastically reduces the cost of adding smart features to new products, which means that people may not stop to think about the potential biases of such a system and their potential for misuse.
For example, Amazon was forced to scrap a machine learning tool used to screen CVs which turned out to be unintentionally biased against female applicants.
You recently had a fundraising round, how is that money going to be used to drive the business forward?
We had three goals for raising capital. First, geographic expansion - we have clients in France, Belgium, Russia and the UK and are looking to expand further across Europe and into the US.
Product development was another goal - we are looking to further develop the platform and launch a SaaS version for the mass market. And thirdly, industry expansion - we want to use the capital to expand beyond financial services, into Telecom, E-commerce and the non-profit sectors, among others.