“AI will be a layer across all facets of tech & business, much like the internet is now.” Rudina Seseri
This past week, the ZyloTeam attended AI World in Boston. We saw robots and chatbots, hardware and software. Companies from all around came to show off the latest and greatest AI technologies. In case you weren’t able to attend, we’ve compiled a roundup of some of the more prevalent themes at the conference.
Virtual Personal Assistants/Chatbots
A major theme we saw was the concept of “intelligent assistants”. Whether text or voice, companies are relying more and more on non-human means for things like customer assistance. Using AI, rather than employing humans, allows companies to handle more customer queries, and frees up man power to focus on bigger issues.
The key here is that the AI technologies are able to communicate and understand natural language (i.e. you can speak to it the way you would to a human). This is easier to execute over text than it is using voice recognition, because people expect a lag when interacting over text and there are less less variables that impact understanding (background noise, static etc).
In a session with Jeff Adams, CEO of Cobalt Speech, we dug deeper into the what/how of voice recognition. There are a variety of components when it comes to the success or failure of this technology. Something like voice dialing, for example, is a fairly simple use case. One person is speaking deliberately into the mic, with little background disruption. On the flip side, a device like the Amazon Echo is more complex in that it can cut through reverb and pick up cues that other devices would miss.
There are certainly limitations to what we’re currently able to do with virtual assistants. Adams cited the example of AI transcribing a meeting. It’s near impossible for any voice recognition software to pick up multiple voices in a large room and accurately collect the words they are saying. The best example of this type of transcription only correctly picks up 20% of the conversation.
A big takeaway from the conference was the importance of creating a branded bot/virtual assistant. You don’t need to make the next Siri/Alexa, but you should create some kind of tech that works towards your company-specific goals.
Image Analysis with Qualitative Reporting & Insights
Another key topic at AI World was image recognition, categorization and reporting. Our communication is centering more and more around images. This means that marketers need to adapt, or risk losing critical insights about their customers. There are many potential applications here. Some marketplaces are using AI to curate the right image for each individual customer. Others are using image recognition to find logos in customer photos in order to build brand affinity models.
Customer Data & Insights Platforms
Companies are building automated systems to identify, unify, cleanse, and enrich your data from both 1st & 3rd party sources. Beyond the data curation, smart platforms can use that AI enabled data to power deep insights and predictive / prescriptive analytics.
The average marketer only uses around 15% of available customer data, so unlocking the full data stack and feeding it into an AI application can yield major insights in a fraction of the time that traditional approaches take.
An important benefit here, is that the feedback loop in this type of platform lends itself very well to AI optimization. There is near real time validation of how effective recommendations are and, due to the self adjusting nature of AI, it can quickly validate and improve its recommendations for your next campaigns.
Content Marketing & Targeting Tools
For content marketers, a good editor is indispensable and can make or break a campaign. But what if an AI system could take over some, or all, of the tasks we rely on human editors for? With the major advances in NLP, writing, editing, and targeting tools are smarter than ever.
From building brand personas of your content, to real time editing and suggestions as you write, AI infused content marketing tools are very quickly becoming more than a novelty. Speaking from experience, they probably won't replace a trusted editor quite yet, but they're well on their way.
Ads were the first place marketers and data scientists started to work together as a tight team, and it makes sense that there are a vast number of AI tools built to help businesses more intelligently, and quickly, make complex decisions around big ad data. Gone are the days of excel sheets, and significance calculators. Now marketers can rely on machine learning to do all of the leg work, in a fraction of the time.
Testing and Optimization Tools
The last category for marketers to keep an eye on is testing & optimization (T&O). T&O is a natural progression for ML in marketing, as multivariate testing for a big brands can be very complex. With a good data source, a smart platforms can test and optimize around any number of factors. These platforms can move quickly, and utilize a deep spread of data. It's reasonable to imagine that, in the next 10 years, most savvy companies will be running nearly autonomous platforms that personalize and shift their site for each customer.
AI World brought together a variety of companies, and helped the attendees get a better understanding of the Artificial Intelligence space. Along with the technologies mentioned above, we also saw trends in AI for head hunting/hiring, development platforms, robotics, and augmented reality.