Marketers and sales reps are struggling to inch conversion rates over 2.5 percent with fresh, purpose-built technologies. But their best efforts will fall as flat as straight-to-spam email if they don’t nail the timeless arts of effective closing. Specifically, they must engage prospects in personalized conversations. This was almost impossible to do at scale — until now, thanks to artificial intelligence, according to Aman Naimat (pictured), general manager and senior vice president of technology at Demandbase Inc. 

“We’re trying to build intimacy at scale,” Naimat said during a recent Cubecast from Demandbase headquarters in San Francisco.

Demandbase builds business-to-business software solutions around account-based marketing principles. Its AI-infused customer relationship management doesn’t rely on old standbys like form-fills and email drip campaigns. And mind you, those emails aren’t failing just because everyone is on Slack.

“If the email was really good, and it related to you, it related to what you’re looking for, who you are — then it will be effective,” Naimat said. In other words, know thy customer first; packaging a pitch is less important than landing it in the prospect’s gut.

Demandbase is betting the reinforcement-learning AI that Google used to beat Pokémon Go champs can also beat the current business-to-business conversion average. Reinforcement-learning and natural language processing help the company’s next-generation CRM scale out one-on-one account-based marketing to large demographics.

Naimat recently spoke with George Gilbert (@ggilbert41), host of theCUBE, SiliconANGLE Media’s mobile livestreaming studio. Watch Chapter 1 of our video interview with Aman Naimat below:

CRM worms its way into customers’ brains

Sales pros using CRM technology understand its power to reach more prospects than they could reach with other available means. CRM is the most-used technology provided to sales staff, according to a study done by Cite Research on behalf of SugarCRM Inc. Respondents rated CRM’s efficacy higher than that of laptops and even smartphones.

The study highlights how CRM systems have evolved to focus on customer experience rather than simply help sales persons with data logging. Sixty-two percent of respondents said the priority when selecting a CRM is enhancing customer experience; 38 percent say improving staffers’ performance is most important.

“Though there is work to be done, the results of the survey indicate CRM has turned a corner. It is heading in a direction where it has the capability to tell sales teams far more about the customers than they could possibly know on their own,” said Larry Augustin, chief executive officer of SugarCRM.

Watch Chapter 2 of our video interview with Aman Naimat below:

A customer’s experience is satisfactory to the extent that it is tailored for him or her. Since the ratio of human reps to customers leaves the former pitiably lacking manpower, it’s an area thirsty for machines to chip in. Ideally, AI that makes scarily accurate guesses based on data could provide automated personal consultations to prospects.

Based on survey responses and its economic impact model, International Data Corp. has predicted that AI in tandem with CRM activities will increase global business revenue by $1.1 trillion from 2017 to the end of 2021.

CRM market leader Salesforce.com Inc. has built its AI capabilities set, Einstein, into the Salesforce platform to enable predictions and recommendations based on users’ unique business processes and customer data.

Demandbase’s CRM operates on similar principals, spiking a punch of public, internet and proprietary data with customers’ own business data. “Magic happens when you take private data and combine it with large amounts of public data,” Naimat said. Data from the internet is under-used by those applying advanced analytics to business problems, he believes.

“The internet today is the largest source of human knowledge, and it actually knows more than you could imagine,” he said. Simple search engine queries can provide useful data on searchers and search-term associations. Demandbase uses the Bing application program interface for this purpose.

“I wish Google would give their API, but they don’t so that’s our next best choice,” Naimat stated.

Add in the company’s data on customers, and deep, rich associations emerge — insights that can help hone in on conversions and sales. “We have code that can plug into your website, and then start understanding interactions that your customers are having,” he added.

Watch Chapter 3 of our video interview with Aman Naimat below:

RL sees the futures

Algorithms using machine learning, particularly reinforcement learning, connect the varied data dots together. Reinforcement learning helped Google’s AlphaGo AI beat the world’s best Pokémon Go wizards. It’s power to make winning moves relies on incessant simulations based on all possible configurations in a given scheme. When a simulation can produce a hypothetical desirable outcome, the system takes action and makes that move — whether it be in a game or in a B2B sales portal.

In the case of Demandbase’s CRM, “It looks at all your possible futures, and then it figures out in what possible futures you will be a customer — and then it works backwards to figure out where should it take you next,” Naimat said.

The CRM system also removes some of the limits of natural language processing — ironically — by limiting it. Initially, the company tried out certain open-source and proprietary NLP tools to find that they often bite off more than they can chew, according to Seth Myers, senior data scientist at Demandbase.

They tend to grasp the subtleties of specific business dialogue only lightly, he told theCUBE. This is because they go for breadth (all human language) rather than depth (these pharmaceutical companies just merged). 

“If we say, OK, we only want to be able to infer a competitive relationship between two businesses in a […] document, that becomes a much more solvable and manageable problem,” Myers said.

Watch the complete video interview with Seth Myers below:

Demandbase has algorithms that identify hot-button language in documents. For instance, it has one algorithm devoted to recognizing when “Amazon” appears in a document and whether it’s the company or the river.

Without deanonymizing individuals, Demandbase can guess things about IP addressees and people behind them based on what content they’ve searched for, etc. NLP performed on internet traffic data and documents can predict what they are in market for, Myers explained.

This can direct a visitor’s journey through a company’s website and take them to their target much quicker, Naimat pointed out. This is sorely needed in the corporate B2B world, where business people have more pressing matters to attend to than navigating  websites. Demandbase even has a product that can rewrite a website’s content to appeal to individual visitors, Naimat concluded.

Photo: SiliconANGLE