Sales and marketing leaders have long understood the importance of having a clean and comprehensive data set within CRM.  But more and more organizations are now adopting solutions that use automation to capture activity data from email, messaging and other collaboration platforms.  This can be dangerous if not done correctly.

Ingesting sensitive or private information and exposing it in the CRM without intelligent filtering can be catastrophic for your business. For example, accidentally syncing 401Ks of your employees from their Gmail into Salesforce.com can easily lead to a lawsuit. At People.ai we know how important this is and we’ve made sure we are the first vendor to help you prevent sensitive or private information from getting logged into your CRM.

We’re all familiar with sales reps being responsible for the monotonous tasks of entering data into CRM.  It’s required by sales leadership yet leads to a significant loss of useful data because reps simply won’t do it.  People.ai has solved this problem with our Automated Activity Capture capability, thus freeing up sales teams from manual data entry, giving them additional intelligence and ensuring that there is a rich and useful dataset available to help enterprises optimize sales activity.

But not all automated capture is created equal.  

Now that organizations are excited about giving sales teams time back to be more productive and about increasing their selling capacity with the help of CRM Sync, they’re also starting to realize the challenges and risks of such automation (if it’s not done correctly).  The spiraling problem:

  1. CRM is a public system of record inside the enterprise and available to the majority of customer-facing employees
  2. While 80% of customer activity data (email, meetings, phone, and other data, describing your relationship with the customer) should be captured in CRM, about 20% of what’s happening in your sales reps’ inboxes is private and limited only to the participating parties. This includes booking vacations or discussing recent doctor visit with their significant other.  Other employees should not have access to such information.
  3. As automation technologies feed activity data into CRM without filtering, such sensitive or private data may end up being logged in CRM and exposed to other employees who shouldn’t see it.

Without some form of filtering and scanning processes, all personal and sensitive information is potentially exposed in the process of CRM activity capture; resulting in some very uncomfortable “oh-oh” moments and potential lawsuits. And we’ve seen some doozies from other products in the market that do CRM sync blindly, without understanding the nature of the content and roles of the sender and recipient  in the business process:

  • A reps personal 401K email update seen by everyone at the company
  • A sales manager has a critical employee performance email shared
  • Employees know about a reps personal date night
  • A strategic company initiative or an M&A project is being considered and is shared pre-maturely, causing backlash and negative PR cycles when leaked

Privacy First AI Alleviates the Risk

Of course, humans intuitively know which of their own information is sensitive and not appropriate for CRM; that’s likely the only benefit of manual data entry and that’s why Salesforce.com still wants reps to manually enter activities in the CRM (after 20 years of doing it anyway).  However, the loss of activity data and lack of productivity poses a very high cost to the business. Thankfully we finally have the technology that can automate activity capture without creating business exposure and increasing risk. So how do you ensure you get the best of both worlds?  

People.ai is the only AI company that has solved this problem to date.  We have built Sensitive Content Filtering, with multiple patents pending, to identify private and sensitive content within activity data.  During the sync process, any data and records that AI deems to be “sensitive” are redacted before syncing to your CRM. Our Intelligent Matching technology makes sure we only capture activities between people who would be relevant to your business processes (accounts/opportunities, cases, etc). Then, even if the AI decided the activity is business-related, we still analyze and filter email contents in case something slipped through.

People.ai used Natural Language Processing (NLP) to build an engine that learns as more activities are processed, similar to how email spam filtering works. First, we used thousands of known sensitive pieces of content to train a number of machine learning models that classified each activity (those flowing through People.ai systems towards our customers’ CRMs) into 23 sensitive or private information categories across 18 languages. Such categories include Vacation, Medical, Immigration, personally identifiable information (PII), and much more. These models help determine whether a given activity is private, sensitive or not.  We now use these models in a process that evaluates every activity across four steps:

  1. Ingest activities such as email, meetings, etc.
  2. Detect whether an activity is sensitive or not using extensive per-customer rule-sets and machine-learning models that were built.  The outcome of the scoring process will decide whether the content is private, sensitive or not and assign decision confidence score.
  3. Match activities to opportunities and match keywords inactivity content
  4. Redact the contents of a communication (such as email body) if the elements of the activity are tagged as sensitive or private with high confidence. The redacted body will be replaced with “Redacted by People.ai: REASON CODE”.  If the content of the activity is deemed not sensitive, no further action is taken and the activity is synced to the CRM.
  5. Sync to CRM with redacted fields or with all information available as relevant business activity.

Additionally, we enable the customer to add custom filtering for known datasets and criteria that they’d like to add to the rule-sets and models as activities are analyzed. Customers can add various domains and users to the filter lists to ensure certain activities from certain users or even groups of users are redacted or even omitted from syncing to CRM.

This privacy first AI approach protects both employees and the company while ensuring there is clear action on how the various sales activities should be treated – “what goes into CRM and what doesn’t.” Sensitive Content Filtering nearly doubles overall detection and redaction accuracy when compared with prior rule-based iterations. This is increasingly becoming a core tenet of an enterprise-wide solution that provides full visibility into customer activity data across sales.

Why Care?

If your sales teams are still doing manual entry, significant productivity is being lost and as much as 80% of all data is not being captured within CRM. A logical step is to begin looking at AI technologies to automate such tedious manual processes, however, handling of private and sensitive information better be one of the leading criteria.

As I indicated earlier, not all automated capture of contacts and customer activity data is created equal. Many enterprises are realizing the risks of picking any ‘ol automation’ technique and vendor when it’s too late. In fact, we’ve seen one sales acceleration tool vendor doing CRM Sync as a side secondary feature sync over 20,000+ 401K statements of all employees into their CRM. Choose wisely and make sure the solution you choose is enterprise-grade.

Think of this as Insurance; you only care about it when you realize that you needed it but didn’t have it. By then it could be too late. We’ve seen these happen when people are sued because they exposed an employee’s financial benefits or health records to the company, or they face fines when exposing private information and are out of regulatory compliance.

Luckily, People.ai is built with this control in place when you switch us on. For the first time, Automated Activity Capture can be done at a massive scale and without risking a lawsuit. We are de-risking the exposure of private and sensitive content within CRM using the latest AI technologies.  Our approach allows you to effectively address adoption concerns surrounding privacy, gain trust from end users, and leverage more and higher quality data from your go-to-market teams.

To learn more, request a demo today.