By Yakaira Núñez (@DigEthno), Senior Director, Research & Insights, @Salesforce

As a research and insights leader, I have seen my fair share of customer and employee insights help companies build amazing experiences. Conversely, I’ve also seen times when ignoring those insights resulted in experiences that missed the mark. With everything that’s happened over the past year, the consequences for missing the mark have never been higher. As we learned in the first article, building better experiences through automation is mandatory not just to stay relevant—but to do the right thing. In this case, the right thing for a diverse user community and for the business. (And no, those aren’t mutually exclusive.) 

One common mistake is to prioritize metrics and processes over people. While this focuses on efficiency and productivity, it leaves out the human experience. And if we leave out the human experience, that’s when automation investments can go sideways. 

While the current environment calls for rapid change and innovation, that doesn’t mean we should “move fast and break things.” In fact, it’s our responsibility not to. We must be thoughtful and encourage diverse voices to be a part of the conversation. It takes a measured approach centered on helping humans thrive to get automation right, and get it right quickly.

One way to do that is to plan an automation strategy that centers around employee and customer happiness. These four steps can help you identify where automation can partner with—not replace—humans.

Step 1: Capture a diverse set of perspectives

Always start with the human perspective. Don’t rush to optimize people. Instead, emphasize the value of creating quality processes that include diverse perspectives, and gather input from everyone—employees and customers.

Use the following questions to help identify who those individuals are, and where automation may result in bias and unintended harm.

• Who are the primary individuals whose roles would be impacted? 
• Who are the “accidental tourists”—individuals you may not have targeted, but who may end up using features in unexpected ways and what does that unintended use look like?
• Do these individuals represent a diverse group—diverse in the broad sense of race, socioeconomic status, neurodiversity, age, gender, technology aptitude, accessibility, etc.? 

Once you’ve explored who will be impacted, conduct interviews to find out what they want and need. For customers, you might ask: Where do you spend the most time trying to get answers or support? What would make that easier? Where do you feel most frustrated in the process? 

For employees, this might mean asking questions like: Who do you work with to get your job done? Who do you actually like working with? Who do you prefer not to work with? And most importantly: Why? Maybe the challenge is timelines that aren’t aligned, or disjointed processes, or always stepping on each other’s toes to get the job done. When you zero in on the “why,” that’s when things get fun—because that’s your reason to make people’s lives easier and better through automation.  

By bringing customers and employees into the automation conversation, you co-create a solution that gives users a voice, agency, and power in the process, all of which lays the groundwork to facilitate adoption.  

Step 2: Examine existing data to understand the processes and people

The data you collect will help pinpoint where to use automation in partnership with humans, and help build new ways of getting the job done—in other words, workflows—that work for the largest population. But you have to make sure your dataset reflects the end-to-end human experience. This may include telemetry usage data, customer surveys, support tickets, and interviews at scale with different target customers (or outliers). 

Assessing the full spectrum of Net Promoter Score (NPS) and Customer Satisfaction (CSAT) scores is also useful. What’s working well for the promoters and how can you carry that forward? What processes are failing for the detractors and how can you make it better? 

And, just as it’s critical to capture the voice of a diverse user group, you also need to be accountable and curate data from a diverse datasets. In order to do so you need to be able to ask and address core questions:

• Is the data you’re collecting truly representative of the population that might use or be impacted by the product? 
• Do you have diverse perspectives analyzing the data for insights? 
• How are you managing for the inevitable (often unconscious) bias you bring to the analysis?

Once you have the perspectives and data, you can then focus on finding solutions. 

Step 3: Look for pain points and opportunities

A pain point and an opportunity are two sides of the same coin. Oftentimes, companies use automation to remove a pain point like replacing paper forms with digital forms. And while that’s a good place to start, don’t stop there. There will be times when it’s impossible to remove the pain point entirely, but those might be an opportunity to uncover a new feature or better process. 

For example, let’s say it typically takes an employee six months to get a purchase order approval through finance. While you can certainly optimize the approval process with automation, why not also give the employee intelligent vendor recommendations? It’s just one lever, but focusing on this area of opportunity helps to optimize the approval process and improve the employee’s day-to-day experience.

Creating more space for employees to focus on tasks they’re interested in pays dividends to both the customer and the business. Nearly 9 out of 10 executives at high-performing companies say that better employee experience directly leads to better customer experience. 

Step 4: Create feedback loops to drive trust and user engagement

One step that companies often miss is creating an effective feedback loop. Make sure the input you’ve gathered from users doesn’t get thrown over the fence and disappear. Be accountable by owning the feedback, incorporating it, and serving it back to users: 

You said you wanted X, and this is what we delivered. Is it working? How is it failing? Can we make it even better? 

Whether you accomplish this through in-app feedback, idea exchange forums, or in-depth interviews, the goal is to co-create actionable solutions that work for everyone involved and continually evolve with user needs. The more individuals impacted by automation understand the science and process behind it and have skin in the game, the more they will trust and value that automation. 

Remember: you won’t be able to incorporate every single insight all at once. Focus on those that will have the greatest impact on your business. And if certain insights can’t go live right away, don’t throw them out. Oftentimes, you’ll find you need to put them on the shelf for a bit. Eventually, when the time is right—you can dust them off and make them part of the solution.

To make automation human, be human

Connect with users, ask questions, make employees and customers a key part of the process, and keep the conversations and feedback loops going. I’ve seen many automation initiatives spring from the pressures brought by the pandemic.  In fact, you can learn more about them in the first article of this series, which explains how automation and AI are changing business. With that experience fresh in our minds, imagine what we can all do on the road ahead. We can create intelligent automation that puts the human in the driver’s seat so we can all focus more on the areas that matter most.