‘Artificial Intelligence’ (AI) & ‘Machine Learning’ (ML) products & feature innovations have taken an intense investment focus by the SaaS giants in recent years, with every enterprise Marketing Technology stack including Adobe, Salesforce and IBM investing in developing a bespoke, personality branded AI to drive enabled, intelligent and progressive optimisations of business systems, service and marketing with seemingly little to no input or effort required for organisations.
Potentially the most significantly invested and developed AI feature set across CRM & marketing technology in recent years is Salesforce’s “Einstein“, which has been continually developed and expanded since the initial launch in 2016.
And why shouldn’t it? with over 90% of organisations letting customer experience expectations down and with well over half of all technology transformation initiatives failing, organisations are continually searching for a golden bullet to produce significant outcomes and solve business complexities more effortlessly in the face of the current technology transformation landscape which is paved with intense risk for enterprise organisations and with the probability of little reward.
“Plug and play” is often a term associated with the black box AI features offered by Salesforce’s “Einstein”, Adobe’s “Sensei” and IBM’s “Watson”, but aside from a great buzz word when companies can announce internally that they there marketing & CRM is “AI-Enabled” in marketing efforts, what value do these products give struggling organisations today and how should we be harnessing the current state machine learning capabilities of SaaS marekting technology stacks? The answer; Artificial Intelligence is not a golden bullet, but it is a bullet.
SaaS AI products are not a holistic brain (just yet), but rather a collection of independent functions & features
AI is ‘marketed’ brilliantly with a personality, positioned as a single source of growing machine learning intelligence that will understand all your data across sales, service, marketing, experience and more with little to no configuration required to deliver business value.
However, the reality is that it’s just not the intuitive yet, and the GAP in driving value with AI will always be the ability to harness Machine learning models that look at your organisation’s bespoke customer data points and understand your business objectives, requirements and complexities, which AI features just won’t “plug and play” to achieve.
Take Salesforce Marketing Cloud ‘Einstein’ features as an example.
Salesforce has consistently released new AI enhancements into Marketing Cloud over the past 2 years, (fueled by Amazon Web Services) to build machine learning models on customer engagement behaviour and implicit preferences to provide insights and ultimately aim to help marketer’s chase the mythical ‘right message, right time’ for every customer engagement.
Some of these Marketing Cloud AI features include:
- Models & predicts open/click behaviour and profile audiences into loyalty segments based on an individuals engagement data points.
- Models & predicts the relative “best time” to send to an individual subscriber based on their engagement data points.
- Identifies undersaturation/over-saturation of audiences based on engagement data and aims to answer the age old question for retail marketing – “how many emails can I send to my audience?“
While these product enhancements look amazingly cool and add the ability for marketers to very easily enhance marketing with AI-enabled modelling to a point, they are limited by Salesforce Marekting Cloud system data and basic customer data like email address, open, click, bounce and unsubscribe data, and they lack the fundamental ability to integrate with your unique CRM of customer data points to enhance the modelling or build your own targeted modelling or prediction based on your organisations unique data.
In other words these features won’t be able to understand your business complexities, objectives and data points to give you customised AI for your marketing.
Ultimately, AI features sets won’t understand your customer better than you do (or can).
The outcome, even Artificially Intelligence starts with understanding your DATA:
- Limited data points for SaaS AI features means limited outcomes
- Out of the box AI marketing features are generally not designed around your organisations data model
- Out of the box AI is not learning based on your companies complete data but on limited data points
- For marketers, AI wont solve your business & process complexities… just yet
So then, if Salesforce, Adobe & IBM’s AI & machine learning features are a toolset, which needs to be developed & integrated correctly to be harness by skilled marketers and technologists, how can organisations make the most of these innovations today?
- Know the strengths and limits of existing AI features sets and position accordingly for you organisations use cases.
- Understand your data & customer behaviours first, in other words build your data readiness to feed AI and other predictive / machine learning models.
- Utilise AI to enhance & supercharge marketing efforts, don’t aim for AI to replace them.
- Test, optimise & challenge AI features against traditional analytics & propensities of your organisation.
- Lastly, invest into building your own organisational capability for building machine learning modelling, this will enable you to
- Use the SaaS AI features available today to full capacity.
- Blend & build upon SaaS AI with your bespoke business data specific machine learning modelling.
- Ultimately, drive the most value and ROI from the existing AI features sets.