It is one of those corporate rites of passage when a just-about-still-a start up company opens the doors on its first annual conference. AI/search/analytics specialist, ThoughtSpot, used its first time in the spotlight to announce that it is already moving into the world of speech-based natural language querying as the new sharp end user interface for its search-based data analytics tools.
Known as SearchIQ, the new addition was seen by many of the 600-Plus delegates at this first conference as the `star turn’ of additions built into the latest ThoughtSpot Version 5 that emerged into the light of day at `Beyond 2018’, held at the National Convention Center in Washington DC.
The conference highlighted an interesting trend within the company, where the focus of attention is already moving beyond the technology of using search-based operations, coupled with significant amounts of machine learning and artificial intelligence capabilities and on to the ways in which a much greater number of individuals can access relevant data, ask their own questions of it and get answers in near-real time.
As joint founder and Executive Chair, Ajeet Singh, told the company’s first-ever keynote session:
Data is not just a visualisation problem, it is now a human-scale user experience problem. The need now is to produce a result for every human being that needs it, not just the data experts. And the world deserves more than what is out there for them to use. What is needed is less input and more output. LIMO will be the company’s fundamental theme for some time to come.
This should not be taken as an implied threat to the future existence of data scientists, but is an acceptance by the company’s senior management that they are in real danger of becoming the next big bottleneck that businesses will have to circumnavigate. There will be much for them still to do at the high end of delivering data analytics and modelling. But for many (still complex) general business management analytics, the need now is for much faster, more automated and far more interactive tools. Here, data scientists are already starting to get in the way.
The data democratisers
For most business users the alternative is no longer an option, either, as Singh pointed out:
Why does access to data have to be so hard? Opinions are of great value, but they have to be grounded in facts.
The `facts’ for just about every business are nearly always available somewhere, but getting at them in order to exploit them is still the biggest problem. This is where ThoughtSpot then claims that its search-based approach to analytics, coupled with machine learning on the common searches users need to employ and AI to both predict and pre-load the answers to those subsequent most likely search candidates, can now provide a solution. To Singh, this about democratising access to business insights.
As new CEO Sudheesh Nair observed, many businesses are less well-served in this area than the majority of individuals. With smartphones, just about everyone these days has direct and almost immediate access to all the data, information and advice they need in their hand or pocket:
But in business they don’t often have access to all the data they need. People still use their guts and experience to make judgements. So, these days they can turn to data scientists to run some investigations.
But 'data' is the new crude oil. It is difficult to extract, and it is dirty stuff. And like oil, it is absolutely useless in its crude form. It needs refining, but there are still only a few people who know how to do that. So data is not just part of the problem, often it is the problem.
For example, Nair pointed to the imbalance of information that the smartphone has made possible There are some 70 billion phone call made every month, but there are significantly more text messages – 900 billion. In business, many of them will be about customers, their orders, lack of orders, product shipments and the rest, all that data needs to be available to the staff that directly deal with such issues, and not just as a post-hoc historical report to their line manager.
Every customer-facing staff member needs real time access to current, real time data on a customer every a contact is made, for they will need to speak to the individuals that understand their business and the context of any question or discussion points they have, so that answers can be provided in real time rather than hearing the old classic, 'I’ll get back to you on that'.
And using a search basis to its analytical tools means that the analysis is being driven and directed by that individual who has the best knowledge of the customer, in real time, rather than by someone referencing an historical, out of date (and relevant data) report. Nair noted:
Search is not a side-hustle. It is the foundations of a building, not the roof or a window. The need now is to democratise access to data and the insights that follow. These days, users need answers to the profoundest question. 'tell me something I should know’. Get it to provide a list ranked by priority and get the human to re-rank them, that is the type if interaction in real time they have aimed at producing.
Boss calling data -what’s happening?
It is easy to assume that SearchIQ is something of a gimmick but initial responses from delegates at the conference suggest that this could be a tool of its time. Given the company’s LIMO objectives, coupled with the desire to democratise the availability of valuable information and insights, it is then interesting to hear delegates talking in terms of applying such notions to senior management as well as front-line troops.
Follow-up features will explore this potential some more, but the notion of senior management being able to interrogate business data by speech recognition and synthesis – effectively converse with it – while driving to work or similar circumstances is expected to prove particularly popular. In the words of one of the company’s little mantras, the issue here is the two-headed question: what is the cost of asking the questions, but also what is the cost of NOT asking them?
The company also quotes surveys from both Gartner and the Harvard Business Review which suggest, respectively, that 38% of organisations are planning to implement or are already actively experimenting with conversational interfaces and voice technology, and that 63% of workers want more opportunities for self-service access to critical knowledge and data.
SpeechIQ works together with SpotIQ, the company’s self-driving analytics engine. It can be set to do tasks and then left to get on with it, such as tracking sales by product and region, both historically and going forward. In this way it is providing required reports out into the future, as well as building growing historical trends records. It can also be set to analyse what happened when a change occurs and report on it.
The system uses machine learning tools to build knowledge of such processes from the data it has gathered and from staff responses to the results. And the company claims it can do this for thousands of users simultaneously.
The latest version of this, launched with ThoughtSpot V5, is SpotIQ AMP. Developed in collaboration with DataRobot to integrate closely with the latter’s automated machine learning platform, this moves the company into the predictive analytics field. DataRobot provides an environment in which a wide range of AI and machine learning algorithms are available to be selected, based on their appropriateness to a specific task.
It will be available as a service from a wide range of cloud service providers, and the fact that it has been available for some time to run on VMware systems on premise now means, following VMware’s recent announcements, that it can also be available on any cloud service that VMware has signed up. ThoughtSpot already provides support for Amazon Web Services and Microsoft Azure, and these two have now been joined by Google Cloud Platform, with its certification to run on that cloud service announced at the conference.
It is still early days to suggest that ThoughtSpot will single-handedly define the next stage of development in the relationship between business and the 'T' of IT. But its use of that 'T' to mask as much of the impact and importance of `T’ is, I feel an important step. It creates a level of abstraction between the 'T' and why anyone would want to use it.
The notion of democratising the insights of analytics has to be the right direction. Putting the insights directly into the hands of the coalface workers, and giving them the chance to modify the analyses to actually fit the work they do, just has to be the right way to go. And it has to be right to attempt to break through next year’s bottleneck of the elite, the hegemony of a small band of data scientists, before it really gets in the way.