The buzz around machine learning (ML) is only going to get louder in 2019 and beyond.
Sure, many of the big names have been harnessing the technology for a while. Uber has Michelangelo, scaling from 0 to 100+ ML models in three years. At Facebook, ML is described as ‘essential’. JP Morgan used ML to complete a task within seconds that normally took 360,000 hours each year.
How about those organisations who have the drive, but face obstacles to adoption?
Deloitte cites 5 vectors of progress towards greater adoption of machine learning in the enterprise: automating data science, reducing need for training data, accelerating training , explaining results, deploying locally
Whether you’re a CMO, CDO or another senior executive focused on delivering great CX, you’ll have heard things like:
“It’s too expensive”
“We don’t have the tech, talent, or the data”
Answering these objections tends to boil down to:
Integrating ML… and mapping it to internal KPIs.
So start low, something that requires the minimum buy-in and investment.
Identifying new forms of revenue is possible, but may be best saved once you have proved a business case.
That’s why CX is the ideal testing ground. It’s seen as ‘soft’ but has hugely transformative potential.
ML data checklist:
- What CX results are you looking to gain?(if you’re starting lean, this can be less pivotal)
At this stage ask if ML is even the right solution for your CX. The PIE framework is ideal for this:
Potential (How much improvement in CX can you get from this particular area)
Importance (how important is it to your organisation, does it directly impact your CX)
Ease (how easy will it be set this up)
- What needs to happen to get this result?
What customer experience does your company need to deliver?
- Where is the data located for your ML?
Is it accessible, and can you use it (legally). And of course, is it clean, non-duplicated and accurate.
- If the data isn’t available, can you get it?
Find out if you can take a shortcut with your data and free it from your siloes. If not, you’ll need to set up the equivalent of beacons, to capture data at every stage of the customer journey. Decide if you have the in-house expertise within your data science team, or if you need a partner
Turning insight into optimisation: Voice of the customer
Let’s assume you’ve got the data.
Imagine you’re planning to capture the voice of your customer.
Stage 1: Persona data
Start with uncovering how customers interact with your brand. Crawl social media mentions, dive into your CRM, ERP and any other databases you’re using
Stage 2: Crunching
Traditional methods would probably include running AB tests, perhaps different versions of the same page. ML will cluster the data from your personas and serve up a page variation based on that. Even before that, it may identify a previously unknown segment for you to target. Of course, once you start getting data from that segment, the data becomes smarter – and so do your campaigns.
Here’s the thing: You don’t need to be a data scientist to implement ML at your organisation. There are platforms out there that will do that for you. Here’s one example:
Time for some answers…
Want to turbocharge your CX strategy with machine learning?
Try the CX2030 Machine Learning Readiness Assessment.
It’s a tool built specifically for CX leaders. You, in other words.
Answer a mix of yes/no questions (it should take you about five minutes).
CX2030 will crunch your answers and give you a score showing where you’re at, where you want to go, and how best to get there.