Musings from the 4th Intelligent Assistants Conference
Late September, the 4th Intelligent Assistants Conference took place in San Francisco. Over the years, this conference has become a flagship event for the space, combining experience sharing from users and panel discussions with vendors. Presentations have just been posted.
A new market landscape
We took the opportunity of the conference to update our market landscape. We continued to wrestle with weeding out scripted bots and preserve our focus on conversational assistants. Natural Language Processing (NLP) and Natural Language Understanding (NLU) are at the “peak of inflated expectations.” Besides, access to these technologies has also gotten easier with many platforms or open source options. So, it has become easy for vendors to claim such capabilities. Yet, offering a truly conversational solution remains hard.
NLU subtleties and the proliferation of messaging applications have stimulated the development of platforms. We broke down the category to better highlight three core capabilities:
- Platforms offered by messaging applications giving access to them as communication channels,
- Conversational platforms providing NLP capabilities, and
- Development platforms abstracting communication channels and allowing assistants and bots to be developed once and work across many.
We also created a new category, Metabots. These "bots of bots" help navigate an ever-growing number of bots and assistants, playing the role of an "App Store." We expect them not just to direct a user to the most relevant one but orchestrate the back and forth between several that may be useful in a conversation. We are in the first inning of this technology. It carries the promise of becoming the primary way to engage users. Amazon Alexa, Apple Siri, Google Assistant, Microsoft Cortana, and Samsung Bixty provide the initial iteration of these Metabots.
No longer just your customer service assistant
The market, once dominated by mobile personal assistants and virtual customer assistants is getting more diverse with the blossoming of new types of assistance. Many use cases are only represented by a few participants and we had to group them in our landscape. But we captured the diversity of bots in our detailed blueprint.
Sales and marketing have become particularly active. The sales assistant space was pioneered by Conversica and now includes half a dozen players such as Amplemarket or Growbots. Automat and Conversable are spearheading conversational marketing. It allows companies to scale their presence in digital channels. L’Oréal shared its experience with Beauty Gifter, an advisor helping consumers explore beauty and skincare products. Its ultimate goal is to drive engagement with the brand.
From cost saving to strategic impact
For many years, assistants focused on deflecting interactions away from a person to self-service to save costs. Several of the projects discussed at the conference showed a clear evolution. Assistants are now playing a more strategic role, enabling digital engagement and improving the customer experience.
DBS Bank discussed the launch of Digibank. Digibank is the first digital-only bank in India. It operates with no branches and only one-fifth of the resources of a traditional bank. Launched in 2016, it gained a million customers in a year. Digibank assistant was central to its success: it handles over 80% of customer inquiries with only 20% requiring chatting with an agent.
Royal Bank of Canada explained how it modernized its Interactive Voice Response (IVR) using the latest speech recognition and natural language technologies. Introduced earlier this year, its new Credit Card Service Line reached impressive precision levels, 70% for authentication and 93% for understanding. It not only drove up self-service usage but improved customer satisfaction.
The Australian Intellectual Property (IP) Department went one step further. It leveraged its assistant, Alex, to drive the organization's digital transformation. In four years, it grew the use of digital channels from 12 to 99% with an 88% customer satisfaction.
These projects often need to use an agile methodology to be successful. Royal Bank of Canada found half of the complaints it was collecting to be related to issues that could readily be addressed. Hence, continuously tuning the solution is crucial.
Eventually, the rise of the strategic profile of these initiatives comes together with new measures of success. I was thrilled to see the emergence of outcome-related metrics like the speed of resolution of customer issues or the flawless execution of their requests.
Intelligence Assistance has definitely crossed the chasm!