Rethinking Enterprise Search Q&A with Massood Zarrabian CEO of BA Insight, and Sean Coleman, Chief Customer Officer of BA Insight

JW: Massood, how does BA Insight deliver the Google search bar experience?

MZ: There are two parts to it. At the user level, we deliver it by itself, standalone. We support multiple search engines, but what we realized over the past years—and, in fact, this innovation came from customers—is that there are people who just don't like a platform and they won't go to it. There are people who live in a platform and it's easier for them. So we moved away from having the search bar live by itself or live in Office 365. Instead, it lives in the systems people use a lot. We integrated a search bar into Teams, SharePoint on-premise, Office 365, ServiceNow, Dynamics, Outlook, Salesforce, Salesforce portal. Behind that, systems are connected. We have 90 connectors, so that when people search something, they can search in multiple systems, and the result comes back the same way, but they launch it differently.

So that's the user side. On the deployment side, we give customers the option of deploying in their environment, and many deploy in their private cloud, or they can take advantage of our SaaS offering, which uses AWS as a platform.

JW: What types of technologies do you leverage behind the scenes to make this kind of comprehensive information access possible?

MZ: We have a platform which embeds Amazon Open Search Service for our SaaS offering, but we are into flexibility. In the back end, we now support eight search engines. We support Kendra, for example. We support multiple versions of Elastic on-prem and in the cloud. We support Azure. We support Solr. We support SharePoint on-prem and SharePoint Online. So, in the back end, the customers have lots of options. Also, in terms of what people call cognitive services or AI—depending on what marketing jargon you want to use—we support the Amazon platform pretty much across the board—Comprehend, Lex. And then, on the Microsoft side, we support Azure Cognitive Services. And, we have actually developed integration of the cognitive services for Google, but we have not marketed it yet. We are just trying to figure out, from a market-demand perspective, what the customer interest is. We demo it here and there to get feedback, but it's not officially supported. For people who are still very much on-premise, we also support things like Grasso, which is open source and can be installed on-premise, or Speccy. So, we give customers a lot of flexibility on what to use and we actually let them mix and match what they want for the best thing, because there are times that some capability from one vendor is better than something that somebody else has and we want to give them the ability to do what they need to do to get the solution to their users.

SC: I think the only thing that I would add on to what you said, Massood, is that the goal is flexibility. There are visionary technologies out there. We've seen it, because, as Massood said, we've partnered with Amazon and Microsoft. We've seen Microsoft add text analytics for medical to their capability after Amazon had medical. So there's this race—this visionary race. And at the same time, there's a TCO thing that's happening—total cost of ownership in cloud services is going down so it's cheaper, then, to run services in the cloud. Sometimes it's cheaper to run Amazon Open Search. Sometimes the Kendra use case is good, but then Microsoft adds semantic search capability to Azure Cognitive. So, while this race is happening, the key point is you want to be able to mix and match and pick and choose and then have a very low barrier of change.

So, if I'm using one today and then all of a sudden there's a huge advance from another, I shouldn't have to rearchitect my whole solution to take advantage of that or to lower my TCO.

JW: Thanks, Sean. So, what is the implementation time for organizations, and what do they have to do to make this work?

SC: In the old days, people would think, "Oh, these projects are a year I've got to go through and I've got to touch every piece of my organization." And that's just not the case anymore. You can drive value in search in as little as a couple of weeks. Larger implementations might go to 4-6 weeks, or a little bit longer than that. But, really, the rule of thumb is that the implementation team on this isn't huge—maybe it's a fraction of a business person, maybe it's a fraction of an IT person working along with us—and it actually gets a lot easier if you use our SaaS platform.

And then, ongoing, you're only using a quarter of a person to keep an eye on things, and that's running reports and looking at relevancy and things like this. Our approach is always to take phased approaches—take a bite-sized chunk of something and deliver value, and then build on that value that you've built over time.