The image was created with the help of Microsoft Design

Can you imagine that in the near future, large language models will become a major innovation tool?

Sure, why not?

I envision LLMs being increasingly engaged in generating new ideas and solving problems by analyzing large amounts of data and identifying patterns that humans might not be able to see.

If we believe — and I do — that innovation is about connecting the proverbial dots, then what can be better at seeing more connectable dots than an LLM?

But for now, we’re stuck with doing innovation in an old-fashioned way: by interacting with other human beings.

That’s where the role of connecting human “dots” (a.k.a. networking) in the innovation process comes into the spotlight.

Innovation that just dried up

In one of my previous posts, I wrote about a profound negative effect on innovation caused by Prohibition, a nationwide ban on the sale of alcoholic beverages in 1920–1933.

Before the passage of federal prohibition, U.S. states and counties could determine for themselves whether or not to allow alcohol consumption in bars and saloons. When federal prohibition went into effect, counties that were previously wet saw an 8–18% drop in patenting relative to consistently dry counties in the same state.

In a brilliant 2020 paper, “Bar Talk: Informal Social Interactions, Alcohol Prohibition, and Invention,” Michael Andrews shows that the effect of Prohibition on innovation had nothing to do with alcohol consumption per se.

Instead, Prohibition affected innovation by disrupting natural social networks.

Prior to Prohibition, bars and saloons acted as social hubs in which individuals could exchange information in an informal setting. When people stopped going to bars, they had lost a “platform” for exchanging ideas. The damage to the innovation process materialized almost immediately.

Collaborating from home

It’s fascinating to see parallels between Prohibition, an ill-conceived government regulation, and the COVID-19 pandemic, a natural disaster. (I’m not a subscriber to a theory that COVID-19 has been a Frankenbaby born in China’s Wuhan Institute of Virology.)

And yet, both did exactly the same: they disrupted natural human networks, Prohibition by preventing people from socializing after work, and COVID-19 by abrupt shifting to remote work.

A recent paper took a look at what happened to employee collaboration when Microsoft shifted, firm-wide, to remote work during the first six months of 2020.

The paper shows that the shift to remote work caused business groups within Microsoft to become less interconnected, rendering the collaborative networks more static and siloed.

Working from home, Microsoft employees did continue to collaborate with others around the firm — using all available digital communication tools — but the pattern of collaboration changed. They were spending more time working with their strong ties but less time interacting with their weak ties.

This is important because strong ties are better suited for transferring information, whereas weak ties are more likely to provide access to new information.

In other words, remote work led to a pattern of collaboration that is less likely to result in generating novel, potentially innovative ideas.

It must be stressed that the timeframe of the study didn’t allow evaluating the effect of remorse work on innovation per se — only on the pattern of employee collaboration.

But it’d be naive to expect that the disruption of established collaboration networks caused by the abrupt shift to remote work will not cause a long-lasting negative (“Prohibition-style”) effect on innovation at Microsoft — and at other tech firms as well, for that matter.

Back to office, back to “normal”?

Recently, we’ve been seeing increasing attempts by firms to get away from remote work and bring employees back to office — all under the banner of restoring the damaged pattern of employee collaboration.

On the surface of things, this makes sense. The “return-to-office” policy can be viewed as an attempt, whether conscious or not, at rebuilding damaged corporate innovation networks and bringing them, as close as possible, to the pre-pandemic state.

Unfortunately, this pattern of rebuilding — “forceful” is the word that comes to mind first — carries a risk. By and large, corporate employees are unhappy with the mandatory return to offices. According to a survey conducted by Clarify Capital, 68% of employees said they would rather look for a new job than return to the office.

This might be more a threat than real action. But let’s not forget that a powerful driving force of innovation is freedom. Yes, you can force people back to their offices and make them communicate in person. But will these unhappy people freely exchange creative ideas with other unhappy people?

I have my doubts.

Besides, although I fully appreciate the role serendipitous encounters may play in the innovation process — a point many executives invoke when pushing for return to office — I’m not convinced that corporate innovation must so much depend on a chance bumping of people into each other.

Can we innovate at scale while waiting for a spark of magic happening at a watercooler or in a bathroom?

A network that is always on

Of course, I’m all for people meeting their colleagues regularly (which doesn’t necessarily mean every day, does it). But do we really need offices for that?

Why won’t we bring employees to off-site meetings instead?

Placing employees in enjoyable settings — with ample time for physical activity and relaxation — will generate conditions for a free and unrushed flow of ideas between the participants. This is a better chance of them generating innovative ideas than talking to each other over a cubicle wall.

Besides, we need to supplement in-person interactions with a platform that would support employee collaboration regardless of employees’ physical presence in the office.

Such a platform should meet two basic requirements:

• Be always-on. Corporate innovation struggles when run as a series of “events” following each other, often with a long pause between them. Instead, corporate innovation must be a continuous flow of parallel innovation projects with fixed starts and finishes, so that at least a few are always active at any time.
• Be project-based. When it comes to innovation, people should have reasons to collaborate, not just “exchange ideas.” This is best achieved by applying the firm’s innovation potential to solving specific, strategically important to the firm, problems.

There is no need to start from scratch: a prototype for such a platform already exists.

I call it internal innovation networks, and I’ll return to this topic in my next piece.