Feature
By next year, natural language processing and conversational analytics will boost analytics and business intelligence adoption to 50% of employees, up from 35%, according to Gartner. Also by next year, Gartner found organizations that offer users access to a curated catalog of internal and external data “will derive twice as much business value from analytics investments as those that do not.”
The research firm reported those findings in its Magic Quadrant for Insight Engines published Sept. 17 (payment required). Insight engines combine search with AI to surface content and data. The point to these numbers? The need for quality enterprise search tools and strategies is growing and investors are taking note. Lucidworks, which provides enterprise search solutions, this summer announced a $100 million investment, about a year and a half after securing $50 million in funding. And Algolia, which focuses on consumer-facing search, raised $110 million in a Series C funding round Oct. 15.
With predictions of increased enterprise search needs and funding for vendors like Lucidworks, we’ve collected some observations about enterprise search technology and strategies as we move into the next decade.
Vendors Investing in AI: What it Means to You
Forrester noted in its Forrester Wave: Cognitive Search, Q2 2019 released May 29 (payment required) that enterprise search vendors that include AI tech like natural language understanding are emerging ahead of the pack. They refer to vendors’ capabilities like “ingestion intelligence” and “intent intelligence” as important components to an enterprise search tool. It should be noted Forrester researchers said cognitive technology in enterprise search is still in a nascent stage.
What does that mean to enterprise search buyers? According to Forrester Wave authors Mike Gualtieri, Srividya Sridharan and Elizabeth Hoberman, buyers should invest in vendors that understand and enrich enterprise data, predict users intent to boost relevancy and include integrated tools for usage analytics, tuning and app dev. Forrester sees Attivio, Coveo, Elastic, Grazitti Interactive, IBM, IHS Markit, Lucidworks, Micro Focus, Microsoft, Mindbreeze, Sinequa and Squirro as the most significant vendors in the market.
Related Article: How Sticky Is Your Enterprise Search?
Evolution of Enterprise Search
Enterprise search as a use case is a relatively dead market, according to Will Hayes, president and CEO of Lucidworks. “While organizations are still looking to leverage the mountains of data they have created over the years and make information more accessible to their employees, the idea that a vendor can simply provide a connector and a search bar and think that this is something that enterprises will value, let alone invest in, is long gone,” Hayes told CMSWire.
Lucidworks’ customers want to leverage machine learning and content analytics, he said, to enrich data for discovery and recommendations to aid in decision making. “In some cases,” Hayes added, “over 100 years of research and insights have been captured by these organizations. They deploy our technology to provide content classification, organization and tagging as well as bespoke workflows using Fusion App Studio to serve various business functions.”
Hayes noted there are vendors who still “carry the torch of traditional enterprise search” and are a “great fit” for departmental use cases which require that technology. He also cited a vendor that provides more straightforward search and index capabilities for point solutions like Salesforce or ServiceNow. “Our stack,” Hayes added, “is much different in that we have made significant investments in machine learning models over the years and focus on the application workflows and overall experiences that these models drive.”
Does Cognitive Search Really Work?
Martin White, managing director of Intranet Focus, CMSWire contributing author and longtime enterprise search observer, said the enterprise search space can still be “incredibly messy.” Skeptical of cognitive technology as an enterprise search capability, White told CMSWire the vendor promise of cognitive technology surfacing relevant information at the top of the enterprise search results is not yet realistic. “Google can make a pretty good guess of what you want,” White said, “but that's not the case inside the enterprise. It’s very difficult indeed to match to what the context is of the searcher, and all the technology companies keep on saying what you really need is cognitive search to help you understand and get a better search.” However, White said, there is a lack of evidence it actually works, he finds.
Related Article: What Do We Mean By 'Search'?
Learning Opportunities
Webinar
May
9
Driving Customer-Centricity in the Financial Services Industry
Discover how financial organizations are leveraging generative AI technologies.
Register
Conference
May
9
An Irish Roundtable in the Valley: The Economic Impact of AI on the World of Work Stanford 2024
Exploring the Impact of AI On the World of Work with Irish diaspora leaders and tech business champions from the North of Ireland
Register
Revealing Post-Purchase Discoveries
Implementation challenges are also noteworthy for enterprise search buyers, according to Martin. After partnering with a vendor, it dawns on some companies quite soon, that the software fees are only a small portion of the total cost of implementation, White said.
White's breakdown? Probably 30% software and 70% personnel costs. “And with that 70%, some of that may be spent internally if the company has got the right team,” White said. “Or it might be spent with the search vendor themselves if they do their own implementation. Or it might be spent with the systems implementer. Those are some big numbers.”
White cautioned buyers to consider the possibility that once every piece of content is indexed, you could discover the enterprise search tech may not work, by which time the search vendor may have "disappeared off into the sunset." “The reality," White added, "is that it's a combination of everything, but mainly rubbish content. Because content inside the enterprise is not curated. No one writes it to be found. They write it in that 20 minutes they've got between this phone call and that phone call. And the fact they even put a page number on it is a miracle.”
Related Article: Diagnosing Enterprise Search Failures
Large Staff Required for Deployment
How do services work for some enterprise search vendors? Hayes said his company licenses its software as an annual subscription. Its typical deployment services, many delivered through Lucidworks' global partner network, run at an average at about 25% of the initial license cost. “We do offer a variety of advisory services for post production enhancement such as relevancy tuning, machine learning training and personalization which can also be purchased as annual subscriptions,” Hayes added.
Implementation remains challenging for vendors, though. Gartner researchers Stephen Emmott, Saniye Alaybeyi and Anthony Mullen noted in their Magic Quadrant that more than one vendor’s customers indicated that the deployment/implementation required the largest cohort of staff from both client and vendor/third party. Another enterprise search vendor has no partner network, Gartner noted.
The Microsoft Elephant in the Search Room
Microsoft has a massive share of this market, according to White. Microsoft owns many digital workplace software deployments, and that includes enterprise search naturally. Microsoft offers Microsoft Search, which, according to Gartner, builds on both the Microsoft Graph and SharePoint Search. It also offers Microsoft Azure Search, which is built on Apache Lucene. Enterprise search owners not getting what they want in the Microsoft enterprise search ecosystem will have to convince their companies to invest in a search application on top of Microsoft, with whom they’ve already invested plenty. Turns out, it cost money to deploy Microsoft in the digital workplace.
It's doable for enterprise search vendors to integrate within the Microsoft ecosystem, according to Hayes of Lucidworks. Some of his company’s largest deployments are with the largest customers of Microsoft and Office 365, he said. “While Office 365 does deploy machine learning in some unique and interesting ways it is centered around the Microsoft ecosystem,” Hayes said, “making it difficult to support data sources from databases, content management systems and other systems of record where large enterprises store data.”
Related Article: Looking at the Evolution of Microsoft Search
Tough Mix of Tech, Content, Training
Bottom line with making enterprise search work? The big problems with enterprise search go beyond technology any vendor offers, according to White. Enterprise search is a “very difficult mix” of technology, content, and, above all, training. “This is a complex business," White said. "It's language based. It’s context based. All the search engine vendors say we can do it blindingly fast, and we've got these technologists to help. But there's no one in the company (of the buyer) that understands how to get the best out of the software.”
No one owns enterprise search, White said, which can make the sales process “very long winded.” And it’s not on the IT agenda. “IT,” he said, “by and large does not fundamentally understand the value of enterprise search.”
