Building a company from ground-up is an excellent learning experience. You soon realize, for instance, that hiring a team member means you’re also embracing their families. A family coalesces into a strong support system that enables the person to give their best to the building of a company.

Parents, spouses, children, siblings, partners, significant others — they know and understand us. Our familial bonds form an invisible support system whose role in our productivity can’t be understated.

In the last few years, we’ve realized that the concept of invisible support systems also applies to machines. How so? Well, it takes some explaining, but it’s pretty much like this:

When I write an email these days, a mail app finishes my sentences for me. Somewhere, in some far-off server farm, a database has untold volumes of written sentences — including my previous emails. An application built with artificial intelligence and machine learning (AI/ML) sifts these emails for patterns. Somehow, the app knows how I’d like to write a sentence when I’m only halfway through.

And then the invisible hand writes the rest. Remarkably, it does it so well that I almost feel like sending somebody a thank you note.

That’s serendipity. It doesn’t happen by accident. Somebody designed this moment. They used empathy, discoverability and learning automation to create moments of happy magic when products do exactly what we want. We don’t have to ask. It just knows. And it delivers.

It’s happening in more apps every day. Shopping sites show custom lists of complementary products. News feeds are personalized to reflect the stories we click on most often.

They’re designed for utilitarian purposes but they make room for happy magic. Serendipity makes a connection deep in the user’s brain. It triggers positive emotions like gratitude and amazement.

How does this design happen?

It starts with empathy — truly understanding the user’s situation at every point in the user interface. Real-time data reveals user intent. AI/ML engines analyze the data, using statistical analysis to predict the user’s expectations and desires at just the right time.

Three pillars of designing for serendipity How should we design for serendipity? Before starting, we should spend an inordinate amount of time seeking and grasping the most powerful word in user delight: context. The entire experience is then about creating that happy magic around the context of the user.

“Seek first to understand, then to be understood” –Stephen R Covey

These three pillars are necessary and sufficient to truly understand context:

• Data: You must meticulously collect data from all possible sources within the context and carefully apply the right filter to seek the truth. For this, you must also have the computing muscle, network bandwidth, storage efficiency and programming talent to deliver on the promises of your user experience. It all has to be optimized for performance, efficiency and economy.
• Design: The user experience must make room for happy magic. It could be a widget with related content, that has the full context of the problem and can ask “would you like to do this next” or a list of the three most likely activities the user expects to do next. Everything has to be strategically placed where the user is most likely to need it — at the precise moment.
• Machine Intelligence: Machine learning algorithms must be thoughtfully engineered and painstakingly designed. The recommendations must be carefully crafted with context in mind for the human and the machine to be two sides of the same intelligence coin.

Learning algorithms will make their own decisions, getting smarter (with human feedback) over time. Learning automation is always about statistical likelihoods, not certainties. You don’t want to leave crucial components of your user experience design to chance. You want it to work like you designed it — every time.

Everybody can design for serendipity It’s tempting to think that sophisticated user experiences designed by the likes of Apple and Tesla are out of reach of everyday startups. Well, that’s just wrong.

Open-source development frameworks give every startup the tools to design their own happy magic. We’re using them in our startup. We’d be crazy not to.

The same is true of data science and learning algorithms. These and many more high end computing capabilities and intuitive low code/no-code development tools are bringing more power and development speed to the people every day.

Of course, bringing AI/ML into the user experience raises a lot of questions. How much user data should we collect? How do we keep it secure? How do we monitor the progress of learning algorithms to avoid crossing boundaries of privacy?

Answering these questions requires a thoughtful analysis of what users will expect and what they will not want. Again, this goes back to my point about empathy.

We always have to design for the entirety of the user experience. We can’t introduce things our users never expected.

So, the challenges are many. As are the risks. But the rewards of combining design and empathy to produce serendipity are too great to pass up.

Let human connections be our guide There’s nothing like that point when you’re in a conversation with a family member or a close friend who knows you so well that they finish your sentence for you. Your eyes lock. You smile and laugh in unison.

That’s the moment my email app aspires to when it finishes my sentences. Of course, it’s just a machine. Our devices can’t replicate the happy magic we get from human interactions.

But we can design for serendipity, building machines inspired by the invisible support systems our families provide.

In closing, I’m dedicating this article to the families of our team building our new startup. Their strong support — not always seen, but always felt — makes us able to bring our best to work every day!