All AI-driven CRM platforms are not created equal

Customer relationships are changing in unpredictable ways compared to a year ago, creating unique data-driven challenges for marketers. The pandemic has made digital convenience a high priority with consumers who want a contextually rich, safe customer experience on any mobile device and are willing to switch brands and products to get it. The shift to ecommerce for everything was accompanied by the expectation of having all daily transactions be digital and touchless.

As marketers struggle to decipher how changes in customer data can affect current and future campaigns, they are looking at what AI and machine learning can do to improve customer relationship management. Salesforce Research’s Sixth Annual State of Marketing Report found that 40% of B2B marketing leaders (manager level or higher) and 38% of B2C marketing leaders planned to increase their use of AI in 2020. That is on top of the 35% of B2B and B2C marketing leaders who said they are already using AI, according to the report. Salesforce surveyed 6,950 full-time marketing leaders from B2B (business-to-business), B2C (business-to-consumer), and B2B2C (business-to-business-to-consumer) companies around the world and found that 84% of marketers said they were using AI in 2020, up from 29% in 2018.

CMOs and their marketing teams face the challenge of delivering results that drive revenue even as the markets redefine themselves. B2B marketing leaders rely on AI to improve their customer segmentation and lookalike audience modeling, according to Drift and the Marketing Artificial Intelligence Institute’s 2020 Marketing Leadership Benchmark Report. Other high priorities include personalizing channel experiences, discovering new data insights, driving next-best actions, including offers in real time, automating customer interactions, and personalizing the overall customer journey. It stands to reason that AI-based CRM applications promise to improve existing processes’ speed and efficiency, increase revenue, and help find new services to sell. Too often, however, the apps are providing more process automation than AI-driven results.

Grading AI’s contributions to CRM

The demand for AI-based apps and platforms is extremely high, leading some CRM vendors to overstate the AI capabilities in their applications, which creates more hype across the CRM landscape. For example, process automation is often sold as AI, when what it does is independently perform simple, repetitive actions and tasks. If it can’t learn from datasets and initiate new workflows or ways of doing work, it isn’t AI.

To provide some context, CRM is big business. Gartner’s latest market estimates and forecasts peg the worldwide CRM market at $56.5 billion in 2019, placing it as the largest segment of the enterprise software market, at 11.7% of global software revenue. Software rating site G2 Crowd lists 14 different types of AI-based CRM applications.

One way to evaluate whether the CRM application actually utilizes AI or is just marketing hype is to look at feature areas individually. Each feature area is assigned a grade in the below list based on how much value AI delivers to marketers and the companies relying on these applications.

AI-based sales assistants or bots
Despite an impressive amount of hype in the CRM community about AI-based sales assistants or bots increasing revenue, they are most often used to automate data entry and scheduling tasks and carry out routine sales force automation (SFA) tasks. Sales assistants or bots in their current generation often only rely on CRM datasets, drastically reducing their possible use cases. Bots are also purpose-built for specific tasks and are often process automation engines. Any organization considering these should give the product area a few generations to get a more integrated foundation in place. (Grade: B- / C)

Configure, price, and quote (CPQ)
Dominated by rules- and constraint-based product configuration and process automation engines, CPQ is another area that’s overhyped when it comes to AI. CPQ benefits most from AI when it comes to guided selling and optimizing revenue management. For CPQ to deliver the maximum value it’s capable of, it needs to at least be integrated with an ERP and CRM system. Constraint-based configurators have been around for decades, as have process automation engines, two technologies that have at times been sold as low-end AI. It’s much easier to use product configuration rules from an existing configurator than to train configuration models, which is what a true AI-based configuration requires. (Grade: C- / D)

Cross-sell and up-sell
Often sold as an integrated app within a configure, price, and quote (CPQ) or account-based marketing (ABM) system or platform, cross-sell and up-sell apps have progressed from relatively simple apps that integrate with product catalogs or product data to more advanced rules- and constraint-based scenarios, including AI-based apps that factor in customers’ personalized preferences. Cross-sell and up-sell apps have become the go-to option in ABM to expand sales into existing accounts yet are limited in how much business value they can deliver using AI. Process automation-based apps are sometimes sold as AI-based in this area of CRM. (Grade: B)

Data intelligence solutions for sales
Vendors providing apps in this category move away from contact and company information and toward contextual intelligence using AI and ML. Vendors’ goals in transitioning to contextual intelligence include supporting sales prospects and selling scenarios with real-time data. Like many of the CRM apps mentioned, vendors in this category do not provide their own datasets. They have limited expertise in improving their data quality to get the most value out of this application. (Grade: B-)

Sales predictive analytics (includes lead scoring)
AI’s impact on improving sales predictive analytics is evident in how effective these applications are in guiding sales rep, sales leader, and sales operations decision-making to improve margins and revenue. The best apps in this category are using machine learning to find new insights in account, sales history, and revenue data. Predictive forecasting, pipeline inspection, opportunity, and lead scoring are a few of the many areas where sales predictive analytics’ AI-based capabilities can contribute. (Grade: A+)

Quota planning
AI- and ML-based quota planning apps are sold as part of an integrated sales performance management (SPM) platform or as a standalone product. The majority of apps in this category today support collaboration and workflows to define accurate, optimal sales quotas. The best apps in this category support various mathematical modeling approaches for assigning quotas across an organization. AI and ML algorithms are being used to set optimal quotas that are in turn distributed across an organization. (Grade: B)

What’s limiting AI’s potential to deliver value in CRM today is the lack of consistent, high-quality data. How much of a contribution AI-based apps and platforms make as part of any CRM system is more dependent on the quality and availability of integrated data and less on the features of the app itself.

Marketing organizations are renowned for having data quality problems, as data governance often isn’t a core strength of the department. There can be conflicting data structures, taxonomies, conflicting metatags, and a lack of consistency across all databases. Overcoming all of these obstacles and improving the data’s quality needs to come first, yet this can be a hurdle too high for marketers to clear. But each of these areas has the potential to deliver greater value in CRM once data quality challenges are overcome.