Laurent - stock.adobe.com
No self-respecting enterprise-level CRM platform has failed to implement AI as a core feature in its offerings over the past few years, and rightly so. AI-powered CRM enables enterprises to know their customers better, offering actionable insights and accurate predictive forecasting.
The inclusion of AI in CRM was unavoidable.
The major CRM vendors -- and most of the minor ones -- exist within larger cloud platforms that are rapidly expanding their capabilities in order to stay relevant in the market. They have all implemented AI generally, and they are now working to build their own flavor of AI into their CRM apps. Evaluating each AI-powered CRM platform becomes a matter of understanding each vendor's particular approach to AI and reviewing how it plays out in its CRM offering specifically.
The most popular AI-powered CRM features
Many of the major CRM platforms have AI features in common. That's expected, both because some aspects of the technology must be ubiquitous to be economical, and because some trends in modern business are impossible to ignore. Common AI features include the following:
- Machine learning. This is the state-of-the-art practice in commercial AI; it involves teaching an automated system to behave productively and effectively in changing conditions, based not on programming or explicit rules, but on patterns and inference.
- Predictive analytics. Now indispensable in enterprise planning and customer interaction, this data science focuses decisions and resources on the most effective course of action at all levels, including personal interactions with individual customers.
- Extracting human time and error from processes is now a standard practice in the enterprise. AI makes automation of complex workflows practical.
Beyond these common features, products can vary dramatically as each vendor takes AI in its own direction. Here are some major AI-powered CRM platforms, including strengths and weaknesses of each.
Salesforce's Einstein is, of course, an industry-leading AI presence, garnering lots of press attention and acting as pacesetter in enterprise AI offerings. Functions include:
- ad hoc analysis of big data, even when that data is from a non-Salesforce source;
- sales/service emphasis, based on Salesforce's many years leading that industry;
- Discovery, an easy-to-use pattern-finding app that works without a complex data model;
- user friendliness to non-specialists; and
- effective AI training resources at the user's fingertips.
Einstein also recently integrated with IBM's Watson, a powerhouse alliance that is only beginning to reveal its full potential.
The platform's weaknesses include modest visualization features and limited or unproven utility beyond the sales and marketing domains.
What does all of this mean for Salesforce CRM? The Discovery app itself is a huge boon, as its pattern-finding utility can be applied to customer behaviors, store performance, and campaign performance and trends. Einstein bots are another benefit. These customer/service-specific smartbots are easily user-configured, hardwired directly into Salesforce customer/product data and include sophisticated scripting features. Moreover, Einstein Prediction Builder and Next Best Action put complex analytics and data-driven AI recommendations in the hands of front-line employees with no data science training.
If Salesforce Einstein is out front in the enterprise game, IBM's Watson is right alongside it. Famous from its Jeopardy! success, Watson has a lot more going for it than quick answers to innocuous questions. Strengths include:
- multi-level accessibility;
- fast performance;
- robust natural language querying;
- strong visualizations;
- straightforward setup;
- deep learning capability;
- strong social media integration; and
- many connectors available for integration with other vendors.
The downsides include a lack of real-time streaming analytics -- a severe applications limitation -- and lack of seamless Hadoop integration.
IBM offers Watson's best functionality in its customer-based tools and boasts the following:
- Watson Discovery -- from the Salesforce partnership -- yields strong customer insights;
- natural language processing;
- Campaign Automation -- Watson Assistant for Marketing -- an intuitive design canvas with Watson in the background, supports personalization, campaign-specific insights and reporting;
- a Lightning-based UI dashboard for AI-based CRM support apps; and
- AI customer service functionality that uses the Watson Assistant for support task automation.
Microsoft Cognitive Services
As one might expect, Microsoft's cloud AI differs from all others in many ways -- some good, some bad. Cognitive Services offers a vast analytical functionality by way of tools that enable customers to build that functionality into their apps, workflows and business systems.
Specifically, it includes the following:
- Databricks data aggregations that organize data for rapid and efficient processing for big data analysis, production applications and research;
- natural language processing, easily embedded in apps and processes;
- Power BI, Power Query and Power BI Data Flows for enterprise analytics;
- deep learning functionality -- somewhat limited to text analysis; and
- strong, flexible, easily configured visualizations via Power BI.
The biggest downside is that Cognitive Services offers little of the ad hoc functionality found in Einstein and Watson.
Several Dynamics 365 modules are directly powered by Cognitive Services, including:
- Sales Insights for data-driven customer insights;
- Customer Insights for campaign and sales personalization;
- Market Insights for global and market segment trends, as well as social media analysis;
- Customer Service Insights for proactive customer relationship maintenance inputs; and
- Cognitive Services, powering the platform's easily configured, adaptable virtual agents.
Oracle artificial intelligence
Oracle has been quietly and steadily working on AI for more than 20 years. The company's efforts began as an initiative to apply intelligent automation to back-end Oracle infrastructure maintenance. Two major products emerged from that undertaking:
- Autonomous Database, a supervised learning-based expert system for Oracle's SaaS database environment. Based on log analysis, it facilitates database performance optimization; and
- Adaptive Intelligent Applications, which comprises machine learning-driven functionality in Oracle's ERP platform.
If Oracle has a limitation, it's that its considerable AI resources, developed within such a limited context, have not yet generalized in functionality as much as its competitors. As for its application in Oracle's CX Cloud, AI enables the following capabilities:
- robust predictive analytics;
- AI-driven customer engagement functionality;
- next best actions, after the fashion of Salesforce;
- AI-driven product configurations, specific to customer opportunity;
- usage analytics on team activity;
- AI-powered chatbots; and
- an AI-driven unified customer experience, focused on optimizing brand loyalty.
SAP's Leonardo hasn't gotten nearly as much press as its competitors, but considerable resources have been brought to bear in support of its robust, generalized AI platform, touted as a "digital innovation system." The system includes:
- versatile, hands-off intelligent workflow;
- advanced analytics;
- blockchain integration;
- cloud-centric generalization of business scenarios for rapid exploitation as new solutions;
- an open platform with open standards; and
- easy integration into other cloud infrastructures, such as AWS, Google and Azure.
Less narrow and constrained than Oracle, SAP's Leonardo is friendly to just about any surrounding cloud system. This makes its way into SAP CRM as follows:
- Conversational AI, a create-your-own customer/employee experience enhancement that is fully integrated into other SAP products;
- easy-to-build digital assistants that exploit the Conversational AI feature, able to integrate with tasks and workflows across channels, humanizing CRM interactions;
- natural language processing, facilitating all of the above; and
- SAP intelligent robotic process automation, applicable across a broad range of customer-related apps and processes, improving response time for customers.
Less familiar still is Adobe's Sensei AI, which is somewhat less versatile than the others, but highly effective in its market. It offers some functionality the others do not (or at least not as conveniently), including:
- Anomaly Detection to determine when something unexpected happens;
- Contribution Analysis, a feature for determining why that unexpected something happened;
- Intelligent Alerts to notify the right people when something unexpected happens; and
- Virtual Analyst, a background cross-channel analysis process that identifies "unknown unknowns."
More conventionally, Sensei includes:
- plain-English querying;
- ad hoc analysis tools;
- software development kits and templates -- Tensorflow and Pytorch -- for its machine learning framework; and
- direct connection to Microsoft's Azure environment.
In the Adobe Experience Cloud, these capabilities enable the following:
- AI-driven customer experience, including the Adobe Target personalization engine for fine-tuning customer contacts and campaigns;
- real-time CRM decision support for relevant customer actions;
- passive reporting on how, when and what customers are shopping for; and
- most interesting of all, Target watches customer online behavior in real time and picks up signals for immediate customer experience modification.
This diversity of approaches to enterprise AI is driving an equally diverse array of AI-powered CRM platforms, each with its own positives and negatives. This diversity will only increase as more CRM vendors join the fray in the coming year.