Artificial intelligence has already infused much of our everyday use of technology — from predictive search results, to email filters, to speech recognition, and autocorrect. Aspects of AI — like machine learning, natural language processing, sentiment analysis and others — have independently been in development for years. So why does it feel like we’re on the cusp of a major landslide past AI assistants into full-fledged computer cognition?
Part of it is, no doubt, hype. As we noted in a recent article, the phrase ‘machine learning’ reached the crest of the Gartner Hype Cycle last year, indicating that our expectations of AI may soon encounter the clear-eyed pragmatism of reality. On the other hand, we have voice-activated home assistants, media platforms with smart recommendations, and cars with collision-detection (which will soon drive themselves) while drones fly around the sky. CRMs are rolling out machine learning features and leading platforms have announced integrations with the aforementioned digital home assistants. Why shouldn’t we be excited for the future?
In the past decade we’ve seen incredible growth in computational power, data production, and neural networks. AI functionality already defines the leaders in CRM. Much the way industry tech was defined in the past between analog and digital, then on-premise and cloud, CRMs will also be ‘dumb’ or ‘smart’, utilitarian vs. forward-thinking. Why spend time setting rules for your own automations when another CRM will create them for you? In coming years, I believe all CRM vendors will offer AI functionality, because the ones that don’t will have disappeared, while the largest vendors battle it out with all-inclusive platforms for enterprise clients.
Artificial Intelligence in CRMs is Making Headlines
There’s been a lot in the news about CRM vendors making partnerships, investments, or acquisitions in companies with promising AI capabilities. Salesforce recently announced the extension of its AI platform, Einstein, to its Service Cloud and Sales Cloud; they will partner Einstein’s AI products with IBM’s Watson. Meanwhile, a Salesforce-developed algorithm has produced surprisingly coherent results for long-text summaries, which is a notorious AI challenge. All of this was announced or reported in the past two months — the summary algorithm in the past few days.
Salesforce’s competitors haven’t sat on the sidelines. Microsoft has been investing heavily in machine learning R&D for years; their acquisition of LinkedIn last year aims to lend highly personalized insights to their Dynamics users’ sales tools. CEO Satya Nadella envisions a common data set for specialized cloud applications. Sounds a bit like Salesforce, no? Meanwhile Oracle — which owns NetSuite, enterprise marketing automation provider Eloqua, and offers various product lines of its Oracle CRM — announced within the past month the rollout of Adaptive Intelligent apps that will span their CRM products, ERP and related business solutions.
Business solution providers were already on the road towards all-inclusive platforms with sophisticated marketing, sales, and service solutions. Artificial intelligence is accelerating that trend. It is the competitive advantage that will allow one enterprise vendor to claw customers from the next — while the benefits trickle down via subscription tiers or packages to SMBs. The capabilities are not yet fully matured, so expect to see many more announcements in the months (and years) to come.
Where is CRM AI Heading?
You don’t need to look far beyond the marketing copy of these CRMs for insight into what these platforms intend their AI to accomplish, but a reasonable expectation of AI development is for these applications to perform at or above the level of a human and do it much faster.
After all, these algorithms will have access to far more data than you or I will ever know and have the power to process it nearly instantaneously. Already, CRM AI is able to automate routine tasks like data entry, record creation and notifications, and higher-level tasks like recommendations, rule creation, lead scoring and analysis. What’s still lacking is the ability to add human-level processing and context (though Salesforce’s summarization algorithm indicates we’re getting a lot closer).
Case in point: sentiment analysis, or the ability to understand how someone feels based on textual clues. It’s extremely difficult, as you have to account for cultural differences, linguistic subtleties, context, and the degree of emotion. As humans, we perform this intuitively, but even then we are not entirely in concordance with one another. An AI performing sentiment analysis at the level of a human would therefore still contain inaccuracies.
But data is created at a dizzying pace; many people voluntarily record minute details of their everyday lives. Suppose a hypothetical algorithm had access to such a breathtaking quantity of data. At the rate of AI development today — with access to nearly a lifetime of context on individual people’s experiences, behavior and personalities — could AI not produce superior and more accurate sentiment analysis than the person reading anonymous text on a screen? I think it’s quite possible (what are privacy laws?).
Who knows? Maybe I’m still stuck in 2016’s hype cycle (or I’m off the sci-fi deep end). Nevertheless, it seems logical that further improvements in AI will reveal that humans are much more measurable and predictable than we let on — even down to a person’s emotions.
Here’s the kicker: is it manipulative to account for someone’s emotional state when determining how and when to induce them to close a sale? That’s an ethics question for another time.
Need a CRM? Check out our free CRM comparison guide.