Salesforce has added extensive AI capabilities (collectively we’ll refer to it as “Salesforce AI”) to its platform. The most obvious and newest capability is Salesforce Einstein, with its generative AI capabilities. But Salesforce has been adding AI to its platform for years before Einstein came along. For example, in 2014 SalesForce acquired RelateIQ. RelateIQ captures data from a range of workplace sources – email, calendar invites, phone calls, social media posts, etc. It then uses AI to analyze the data and provide actionable insights without having a salesperson “think about it” such as a reminder to follow up with a lead in a certain timeframe – with the timeframe being set by the AI based on a predictive model from the data.
But if Salesforce has been becoming an “AI-first” company since 2014 (to quote Marc Benioff’s statement at a company all-hands in 2014), what about all the third-party applications that run on the Salesforce platform? Have they invested in AI to keep up with Salesforce’s continually extending integration with AI? In many ways, it doesn’t matter what Salesforce AI can do if the third-party apps that provide the “last mile” to reaching and communicating with the customer in a personalized way are still lacking in AI capabilities. What good does it do to have fiber optic cable with its high bandwidth if your last mile to the house is still copper? This analogy is quite relevant to Salesforce AI as your entire customer-centered platform must be AI-driven to take full advantage of the capabilities of Salesforce AI.
In this article, we are going to focus on AI for data deduplication, cleaning and validation, which we will refer to collectively as “data integrity” moving forward, since that is what DataGroomr does. We will focus on other aspects of Salesforce AI, including Einstein, in a series of follow-on posts.
AI for Data Integrity
So why should you care about the latest AI capabilities for data integrity in third-party apps along with the capabilities of Salesforce AI? After all, you clean, deduplicate, and validate your data today, so how does AI help or matter? There are several reasons. AI for data integrity has several impacts on you and your organization when combined with Salesforce AI:
- Master Data Management. There is a new area that has developed around machine learning, AI, and analytics called dataOps. In this world, all data flows from a variety of structured, semi-structured and unstructured sources into a central repository called a data lakehouse. This market and functionality are growing rapidly. Soon all data, including Salesforce data, will flow into these central corporate repositories for use in business analytics and machine learning functions. This final data set is called a “golden master” and the process to getting to the cleanest data in the repository is a subset of dataOps functionality called Master Data Management. If you are a Salesforce admin, you will be expected to only send clean data into the central repository. This will become a metric in your KPIs and a core expectation of your role. If the data is dirty, it will be very visible to others in your organization. And you will hear about it if you aren’t doing a good job providing the cleanest data possible.
- Accuracy of Data. This goes along with the first point and is an obvious good in its own right. Why else spend so much time and money on data integrity? But accurate data is not only important for Master Data Management. Salesforce AI’s functionalities also make use of this data. If it isn’t accurate, your Salesforce functionality will be less than optimal and your sales results will suffer.
- Orchestration. Maintaining data integrity is a time-consuming and repetitive process. Most apps performing data integrity functionality today are rules-based, which means you have to spend a great deal of time conceiving the rules, then adding them to the system, validating them, and then running them. And this isn’t a one-time process. You have to do this on a regular basis, over-and-over again. This is an obvious place where Salesforce AI and AI specifically for data integrity can hugely increase the efficiency of what is a mundane, painful task and free up time for things you prefer to do. And not just time at work. Maybe you could use that free time to achieve more personal goals, like spending more time with the family (one of our favorites).
There is a term for when software automates a multistep, repetitive process. It is called orchestration. Orchestration is one of the hottest areas in software right now. IBM’s core Watson platform is highly centered on improving efficiency by using AI for orchestration. Any number of other integration platforms – Workato, Jitterbit, tray.io are making millions solving orchestration problems. And a lot of the Salesforce AI that is going on “under the covers’, such as the example cited above, are all focused on using orchestration to improve efficiency.
- Inaccurate Predictive Models Can Cost Millions in Wasted Marketing Spend. We mentioned the need for accurate predictive models. A great deal of predictive modeling that goes on in corporations today is built on customer data – such as audience segmentation. When the models are done, they spit out audiences by metropolitan statistical areas (MSA) or zip code, which is then sent to advertising agencies and others to send marketing messages by email, direct mail, or online. Companies spend billions of dollars on this every year. Bad customer data will ultimately lead to poor performance of multi-million dollar marketing campaigns.
So that is why data integrity with Salesforce AI is so critically important to solving a key aspect of the “last mile” problem. When you talk to your data integrity vendor, make sure you ask for a demonstration of how they are using AI to handle the critical functions we just discussed. That way you can assure that you are getting the most out of Salesforce AI, maximizing your efficiency, and getting the highest return on your marketing investments.
