Q&A with Michael Schrage

Thinking Beyond Software Agents to ‘Multiple Selves’

Just as the world is falling in love with Siri, Cortana, Alexa, and other software agent/digital assistants that ease life’s pesky daily chores, author, IDE fellow, and digital innovation pundit, Michael Schrage, has a better idea: Why not build digital versions of ourselves that amplify our best human aspects and attributes? These “multiple selves,” he contends, will yield more productive employees, more empathetic companions, and more creative thinkers — not merely automated attendants.

It’s a provocative concept that could be the logical extension of today’s digital assistants as AI and machine learning take hold. But how would they work and what functions would they assume?

In a recent interview, Michael shared some of his thoughts with MIT IDE editor and content manager, Paula Klein. His full IDE research brief on the subject can be found here, and a blog on the topic appears in Sloan Management Review here.


Q: You seem somewhat skeptical — even cynical — about the future of software agents despite their appeal. Please explain how multiple/digital selves fundamentally differ from the design of bots and digital assistants?

A: It’s simple: Is sustainable personal productivity more likely to come from better agents or better selves? My research argues that self-improvement — using digital technology to craft selves that are smarter, bolder, more creative, more persuasive, more forceful, or kinder — is clearly the superior innovation investment.

Don’t get me wrong; digital assistants are terrific. Who wouldn’t want software that anticipates your needs and obeys your commands? I certainly do. But virtual agents like Alexa and Siri only perform tasks; selves actually determine and define them. That distinction matters.

Serious managers — those who truly care about personal and professional development — will look beyond the instant gratification agents provide to consider deeper productivity gains. To me, better agents are tactical; better selves are strategic. It’s strategic to design and build better selves that amplify our best aspects and attributes. In the future, we’ll still have smarter, obedient agents like Siri and Cortana on call, but our better selves will be telling them what to do.

We are in the midst of a technological transformation that can take us from a self-improvement sensibility to a larger, selves-improvement ecosystem. Wouldn’t you be much more effective with a coach or adviser who reliably presents us with data-driven workplace choices and recommendations to bring out our bolder, more creative, more persuasive, more commanding, more empathic selves? Mediocre performance is no longer sustainable in our hyper-competitive business environment.

Q: What prompted this focus on productive multiple selves?

A: Over the past 20 years, my design research emphasis has radically flipped from “How can people create more valuable innovation?” to “How can innovation create more valuable people?” My curiosity about digitally designed and data-driven multiple selves emerged from that.

I’ve always been interested in self-improvement, good advice, and better professional choices, but like a lot of people, I’m short on time and patience. I’ll embrace anything simple, fast, and easy that makes me smarter, more self-aware, and more productive. That’s why I love Google and Bing, self-quantification, and recommendation engines. They’re super-fast, super-convenient, and they consistently alert me to choices I didn’t know I had. For me, that’s empowerment. But, frankly, they feel like pieces in a puzzle that no one has bothered to put together.

Q: How, specifically, did that lead to the multiple selves concept?

A: The idea of multiple selves offers a powerful ‘unit of analysis’ for inspiring a personal productivity design revolution. The conceptual breakthrough emerged from a commingling of relevant economics, behavioral economics, cognitive psychology, and sociological research. The literature in these areas is awesome and compelling.

In essence, the more creatively, comprehensively, and innovatively we digitize our “selves,” the greater the opportunities we have for helping people define, design, and develop the optimal, traits and attributes they desire.

Think, for example, of a portfolio of ‘Pareto Selves’ — that is, what are those 20% chunks of talent, temperament, and behaviors that account for 80% of our impact, influence and value? It would be awesome if we could count on and collaborate with our Pareto Personae at work. This will become a high-performance norm sooner than we expect.

Individuals can choose customized software, digital tools, and user-specific options to improve performance and productivity by addressing key attributes — be they affective qualities like boldness or friendliness, or technical skills such as facilitation and formulating testable hypotheses.

Q: What might an economic/business model for multiple-self workers look like?

In theory, a quality Multiple Selves as a Service provider could charge people $199 per year per self, or 1% of gross income — whatever is less — to sign on. People and their employers can quickly decide if ‘better selves’ are worth the money compared to say, an Amazon Prime, Alexa or Netflix subscription!

More seriously, the real challenge has less to do with business models or technology than data governance. How should companies facilitate data access and sharing of personnel data and workplace analytics? Does your ‘boldly creative self’ at work belong to your employer or to you? How portable or proprietary should one’s data-driven productive selves be?

Everyone with a mobile phone remembers BYOD and apps. Perhaps we move into “BYOMS” — Bring Your Own Multiple Selves — environments where talented employees hack their ‘selves-ware’ into enterprise IT platforms. I don’t know; it’s too early to know how it will unfold. But if I’m Microsoft — which already owns Outlook and LinkedIn — or Salesforce, IBM, Oracle, Google, Slack, AWS, or Alibaba, I’m going to be really interested in how multiple selves in the enterprise radically expand my client opportunities. Do the math: a business employing 200+ people may soon need analytics and support for over 1,000 selves. Global giants like GE, Toyota, WPP, or Randstad NV may productively manage millions! That’s a market.

These speculative scenarios aside, I can’t overstress that — technically — all the data, algorithms, analytics, devices, and almost all the UXs we’d need to create custom multiple selves already exist. Right now, the enterprise focus is on bots, agents, machine learning, AI, wearables, and other point solutions for boosting enterprise productivity. The workday is near when monitors and personal KPI dashboards will physiologically sense when users are not in the mood to take advice. Individuals and their organizations will be able to track which “selves” deliver the best performance and outcomes.

As smart organizations and high-performance talent look for innovative ways to make themselves more valuable, they’ll turn to multiple selves as a powerful organizing principle to make all those other technologies more valuable, as well


Michael Schrage is a research fellow at the MIT Initiative on the Digital Economy (IDE) and a visiting fellow at the Imperial College Department of Innovation and Entrepreneurship in London. He is the author of The Innovator’s Hypothesis: How Cheap Experiments Are Worth More Than Good Ideas (MIT Press, 2014).