#IntelligentAgent Makes B2B Marketing an On-the-Go Reality

Business-to-business (B2B) marketing can be a complex dance. The complicated sales process requires the approval of multiple stakeholders, and contracts can be massive tomes of confusion. Closing a sale can take weeks, months, or even years and involve thousands of touchpoints throughout the process. It can be daunting just to keep track of all the key players and decision-makers as each prospective customer’s business grows, evolves, and adapts to an ever-changing marketplace.

“Most enterprise businesses have invested heavily in sales and engagement software, but there’s a lot of room for improvement. We wanted to see if we could leverage technologies from Adobe and Microsoft to make the B2B sales process more seamless, engaging, and fast,” says Saurabh Khurana, a product marketing leader at Adobe.

The result of that collaboration is a preview technology called #IntelligentAgent, that Saurabh demonstrated on stage at Adobe Summit 2019. #IntelligentAgent is a collection of several different technologies from Adobe and Microsoft, working together.

In the first part of his demonstration, Saurabh showed technology that the team is exploring in Marketo Engage — Adobe’s industry’s leading B2B marketing automation solution — that will make it easier to identify and engage stakeholders in the B2B buying process.

#IntelligentAgent shows how machine learning can be used with Marketo to predictively identify new members of a client’s buying team.

The new technologies in Marketo Engage will provide a real-time view of your clients’ account activity, such as social media posts and business news. It will also use machine learning to analyze Marketo Engage account data, alongside client data from Microsoft Dynamics 365 and other third-party business vendors, to predictively identify key stakeholders and decision-makers who are part of the client’s organization.

“There are a lot of people at the customer’s organization that you need to engage with, and this gives you the competitive advantage of a head start on finding and engaging with critical stakeholders,” Saurabh said.

The second part of #IntelligentAgent makes it easy to extract, consume, and understand critical information from complex business documents — such as contracts — on the go.

Broadly speaking, Saurabh and his team wanted to see how AI and machine learning could be used in practical ways to collect better business intelligence and improve document experiences. “In our brainstorming meetings with Microsoft, we realized that we had complimentary technologies we could apply to gain better insight from documents, personalize documents for unique business situations, and make it easier to consume and understand documents,” he said.

During the demo, Saurabh showed how document analysis technologies can be paired with voice-to-text recognition from Microsoft Azure Cognitive Services inside Adobe Acrobat Reader DC to intelligently extract and summarize key information from complex documents.

Combining a mobile device with a natural, conversational voice query like, “How much does this cost?” #IntelligentAgent can search a contract, identify the relevant price information and display it, along with the supporting charts and data used to extrapolate the answer.

This sneak presentation of #IntelligentAgent shows how price information can be extracted and compared across multiple documents, and then summarized for viewing in a mobile device, using a few simple voice queries.

A more complex question like, “How does this price compare to the previous contract?” can also be explored, and #IntelligentAgent will seek information across all the previous contracts stored in Adobe Document Cloud to find a relevant answer.

“At a high level, we’re using document analysis technology to understand the structure of a document — the header, sections, and object types inside the document — along with natural language-processing AI to get a semantic understanding of what’s in the text and graphs to generate an answer,” Saurabh said.

That can be a real challenge because contractual language is significantly different from conversational language. It’s one of the reasons that #IntelligentAgent was designed — not just to provide an answer, but to highlight the information and details that are used as the basis of that answer.

“One of the tenants of good AI is that it shouldn’t just be a black box that spits out answers,” Saurabh said. “People need to be able to trust that the AI got it right, especially if you’re talking about a contract worth millions of dollars. By highlighting key parts of the underlying data, we’re improving confidence in the result and also giving the user the opportunity to take control and dig deeper into the document, without having to dig through pages and pages of extraneous data or legalese.”

Although some of the technology inside #IntelligentAgent is just prototype, Saurabh noted that since it works inside Acrobat Reader, it has the potential to scale very quickly if it is fully developed and released. “That’s what the Adobe team is all about,” he said. “How do we help customers go from paper to digital, and how can we make those digital experiences more personalized, contextual, and easy to consume? With #IntelligentAgent, we can do that for B2B marketing experiences, accelerating the buying cycle by starting with engagement and ending with signed contracts.”