I’m attending IBM’s Connect event in Florida, and in a discussion with some of the product folks working on Verse, IBM’s intelligent email client, I had a few thoughts. These are equally relevant to the folks working on Inbox at Google and Office365 at Microsoft, however, or those working on intelligent agents in whatever sorts of communications and coworking tools.
A great deal of attention is being spent on what I think of as first-order benefits from smartening up tools. In this example I consider new email clients, and applications such as scheduling meetings more efficiently, bringing the appropriate documents into the context of some exchange with others, or prioritizing emails based on the identity of the senders. That’s all well and good.
I’d like the researchers, developers, and product folks to think about what happens when all the people sending and receiving these emails are using intelligent tools, not just one of them.
However, I think the real bang is hiding in second-order effects, about which I hear little.
By second-order effects I want to draw attention to one class of benefits, in distinction to other possible interpretations of ‘second-order’. I’d like the researchers, developers, and product folks to think about what happens when all the people sending and receiving these emails are using intelligent tools, not just one of them. Not just the persona in the individual use case. Imagine instead what the experience might be when my intelligent email client can be coordinating, cooperating, and communicating with your email client to our mutual benefit.
The obvious simple example is scheduling a meeting. Today when I suggest a meeting in an email, I cc email@example.com, an AI virtual assistant I’ve been using for months, and she then coordinates with the email recipient or recipients, freeing me from the headaches involved. That’s great, and the tool saves me a great deal of time, personally. But in general, all the recipients still have to do the work of coordinating the meeting with Amy. They have to look at their calendars, suggest times, cycle, etc.
The social lives of our ‘bots will soon be the buzzing substratum of our work, one that will simplify and accelerate our efforts. That’s when we will hear the sonic boom coming from the armies of smart agents working with us, for us, under our feet.
In principle, if they are also using x.ai’s tool, they can also get the time savings that I do. And in the rare circumstance that all are using x.ai, instead of me alone avoiding that cascade of emails, we all would. That’s the sort of second-order effect I am looking for.
Another example, one with more sizzle, might be when these agents might know a great deal about us, for example our psychological profiles. My future agent might know that I am an Extraverted iNtuitive Thinking Perceiver (borderline INTP) in the Miers Briggs schema, and the other agent knows that the recipient of an email I am writing is an Extraverted Sensing Thinking Judger. The agents might communicate about the email’s effectiveness given the differences in personality and might negotiate a way to restructure the content — while I am still writing it — to increase the likelihood that my point is better made. For example, toning down my enthusiasm for several extraneous ideas, and minimizing that into a single supporting paragraph following the central issue, which is really about (what I secretly consider boring old) process planning, which is what (process-oriented) Jane, the recipient, really cares about.
These agents could surface their reasoning — if I wanted to see it — or if I had come to accept their tweaks and marginal restructurings I might simply leave them to autopilot the email, with just a quick glance.
Now multiply by the five other recipients on a similar email going out a day later, once Jane and I have come to an agreement on the process under discussion. The conclave of agents may have recommended that instead of an email Jane and I should invite all to a video meeting, and the agents created a deck for the call — based on the email exchange — so that Jane and I could discuss the new process. And they genned up a graphic of the process, since most of the others are visually oriented, not textual.
The social lives of our ‘bots will soon be the buzzing substratum of our work, one that will simplify and accelerate our efforts. That’s when we will hear the sonic boom coming from the armies of smart agents working with us, for us, under our feet. Just as I have come to take modern communications tools and their impacts for granted, as a given, as a basic principle upon which everything else rests, soon we will do the same for the output of these networked agents, weaving the fabric of second-order productivity.
Originally published at stoweboyd.com on February 1 2016.