The use of AI is no longer a competitive advantage, it’s a non-negotiable.
With tools like ChatGPT making AI accessible to the masses and every conversation you have in marketing and sales promising AI will make your life easier, how often do you stop to consider the hidden dangers of what AI produces?
AI is only as good as the prompt and the data it’s fed. If the data is incomplete, you’ll make bad business decisions. If the data is wrong, you’ll miss opportunities. If the data is missing, you can’t trust the answers.
We’ve all adopted AI incredibly quickly, moving from fun prompts for making your likeness as a toy in blister packaging (probably with a MacBook and coffee cup, as per the marketing meta) to integrating AI tools directly into CRM, marketing and sales platforms.
But how often are you stopping to ask whether you can trust the answers AI is providing? Or whether the data being reported looks correct?
There’s no doubt in what AI can help you do. It can help with coding, design, and even simple workflows. It can help you write your next sales outreach email, your next social post, or plan your next work trip. Best of all, it can analyse your data and provide clear answers to help you understand what’s working and what’s not.
But only if the data is ready for AI.
With AI-ready data, you can prompt AI to do things like:
- Identify the marketing channel that contributes to the highest number of closed-won deals and break down key attributes that can be used in future marketing campaigns.
- Provide a list of companies where associated contacts have previously opened an email but haven’t engaged in the last 6 months then create an outreach email to re-engage these contacts using information about the company to personalise the messaging.
- Review the lifecycle stages and provide a breakdown of contacts by lifecycle stage and based on the data, make recommendations for how we can progress more contacts into the sales qualified lead stage.
These are of course just examples, and you might have a list of your own criteria that you really would like AI to take on. Critically of course, prompts like the above rely on having AI connected with your marketing, sales and CRM systems, in order to access the data needed. Furthermore, the data needs to be structured in an AI-ready way to really utilise it.
Following the integration between HubSpot and ChatGPT, HubSpot shared some interesting ideas for prompts, some of which you can access here, but ultimately, it still comes down to data being organised in a way that AI can access, analyse and utilise for reporting, content and operational purposes.
The challenge is AI-ready data
What do we mean by AI-ready? Data that is structured and organised in a way that AI can easily interpret and scrutinise, in order to report on it, provide information on it, or even create content using it.
As well as structure, the completeness of data is vital - just because you have a lot of data, and if you contribute to the customer journey or contribute to revenue, then you probably do have a lot of data, it doesn’t mean it’s AI-ready. Missing information and context can lead to misinterpretation of the data by AI, which can result in incorrect answers and misrepresentation of the truth.
Much like marketing automation platforms or CRM tools, if the data and processes you put into AI aren’t good, then the output won’t be much better.
On the plus side, you can actually use AI to help organise some of that data, for example, by using AI data enrichment, you can broaden the information you have for marketing and sales prospects to enable better use of workflows and personalisation.
AI Scenario:
Automating Customer Lifetime Value (CLTV) reports based on deals in seconds
In this scenario, you have 5 years worth of deal data in your CRM, both open deals and closed. In each deal record, you track the deal amount, close dates and type of service, and all deals are associated with a company.
This alone provides an incredible amount of data to scrutinise and report on, but manually creating and reviewing the reports is time consuming and sometimes a little confusing.
One report in particular that you’re trying to understand is the customer lifetime value (CLTV) of your top 5 customers over the 5 year period as well as the last 12 months. From this comparison, you can better understand if the customers you’ve signed in the last year have higher lifetime values than in the past.
Using AI, you create the following report: Top 5 customers by lifetime value in the past 5 years.
In seconds, the AI tool comes back with the following:
- Company A with CLTV £150,000
- Company B with CLTV £100,000
- Company C with CLTV£100,000
- Company D with CLTV £95,000
- Company E with CLTV £90,000
You’re delighted with the results, which come back quicker than you could even review one of the deals in question. You’re even more delighted that from first glance, the deals have all been closed in the last 18 months, showing a higher CLTV of new deals.
On second glance, you see that the data can’t be right, as you would have expected two of your longest serving customers, company X and Y to be in the list.
Delight quickly becomes frustration as you spend the next hour trying to work out why some customers aren’t in the report. Then it clicks: there is no data on the deal record that shows the length of a deal, only an amount. Worse still, the amounts for retainer work, billed monthly, is very inconsistent, with some deals showing a total for the retainer contract length, and other deals showing the monthly amount as the deal is recurring without an end date.
Immediately this flags that company X should actually be at the top of the list, as despite the deal amount only showing £5,000, this has been billed monthly for 5 years, totalling £300,000.
An even deeper dive reveals that what should be one customer in the CRM is actually represented by 6 different companies in the CRM (some of them with the incorrect spelling of the company name), where sales have created new companies each time they’ve set up a deal, without checking that the company was already set up in the CRM.
The data inefficiencies and bad processes essentially mean that AI can’t correctly report on the data, taking more time to create the reports through AI than it might have taken to do it manually. If the data was inputted properly and the process better matched the reporting needs, then AI would have easily returned the correct answer, but it’s a big ‘if’ and one that demonstrates the need structure and organisation within your CRM in order to fully utilise AI.
In order to get to a point where the data is ready for AI, processes need to be optimised, CRM users need to be disciplined and data from different sources needs to be integrated into one place to allow AI to get the full context of the data. This is where RevOps comes into play.
It’s not just sales data
While the AI scenario focuses on deal data, the underlying message is true for any business function, particularly ones that contribute to revenue and need to report on ROI and financial performance.
Marketing teams generate a huge amount of data, but it’s only usable by AI if it’s tagged correctly and included in the right prompts. How marketing, sales and service teams align themselves is also very important, as cross-function reporting and data usage is common, especially where different departments are trying to better understand the customer journey.
You can’t ask AI to list every customer that has seen a specific page on the website or used a specific function in the app to send a service announcement if you don’t have the data readily accessible in the first place.
Data is the lifeblood of AI and without, functionality will be limited.
RevOps: the route to AI ready data
In this instance, RevOps can be defined as a systematic teardown of every user and customer touchpoint. The set-up can be used to collect the right data, measure and make improvements across sales, marketing and customer success.
Done well, RevOps is a catalyst for growth, clarity, efficiencies, improved retention and better profit margins, and by extension, the route into using AI in any business process or function.
By starting with a RevOps approach to your data, you’ll build a much stronger position for utilising AI. The approach can start in many different ways, but often takes into account your current tech-stack and where the data lies, as well as your current processes.
As a result of properly implementing a RevOps framework in the organisation, you can ensure data is AI ready and make AI do all of the heavy lifting!
Start your RevOps journey today - book a free consultation.
