How NOT to Inspire Change

Apple's recent ad for a new iPad shows what happens when executives fall in love with technology and forget about people. 

woman using ipad

Apple's ability to excite customers about technological change is legendary... but the company erred badly with a recent ad for its new iPad Pro. And if Apple can totally misjudge how people will react to change, then so can you and I. 

I worry, in particular, about how the insurance industry will manage all the changes that are becoming possible with generative AI and that will need to be rolled out throughout companies over the next many years. While executives sing its praises and talk about how much drudgery it can remove from jobs, how much more efficient it can make people and so on, I'm not sure they're fully factoring in the fears that many employees harbor and the organizational changes that will need to occur.

I have thoughts.

So we're all on the same page, here is a link to the Apple ad. And here is a link to an article about Apple's quick apology.

You can see what Apple was trying to do. It wanted to show that a whole array of creative tools — books, musical instruments, a record player, paint and much more — had been combined into a single, sleek iPad. Apple was even being its usual cheeky self by invoking a popular meme, in which people put cans of paint or fruit or just about anything into a metal crusher and then film and share what they look like when they explode.

What Apple somehow missed is that many of the items being crushed are totemic. People love their books, their pianos, their record players. People don't want to see those crushed, even in the name of sleek technological progress. You can't just sell the notion that a technology is cool and assume everyone will climb onboard.

Many people like their jobs, too — even the less efficient parts. They've done the job the same way for a long time, and they're in no hurry to change. Change takes effort and is disruptive mentally. 

A rule of thumb among venture capitalists is that a new product needs to be 10 times better than what a startup is trying to replace, or don't bother. That number doesn't need to be as high with employees because, after all, the employer is paying their salaries. But there still needs to be a clear advantage, or the employee will resist the change, and the employee has to be wooed, not just ordered around.

I've always been a bit of a curmudgeon about change management programs — banners and leaflets and rah-rah meetings aren't my thing — but I thought two senior consultants at Heidrick & Struggles whom I helped with a book eight years ago were quite perceptive on the topic, so I'll share some highlights here. (If you want to investigate further, these points all come from Chapter 11 in "Accelerating Performance.")

Colin Price and Sharon Toye, who have both moved on from Heidrick & Struggles, opened the chapter by emphasizing the need for speed. "When transformations don't work," they write, "the biggest reason is that bold new ideas weren't institutionalized rapidly enough." They then get to the five steps they recommend for an organization trying to make the sorts of changes that generative AI will allow... and demand:

"First, leaders need to connect with their people through a common and compelling purpose. Next, leaders must align the operating model of the organization to reinforce the behavior change. Third, capabilities must be built. There is no point in asking people to do things differently if they don't have the skills to do so. Fourth, the changes must be role-modeled by leaders. We all know that following the parental mantra of do as I say not as I do will ruin any chance of colleagues doing what senior leaders asked of them, although it is surprising how many senior players still try to get away with this manner of leading. Fifth, but by no means last is the need to provide space for people to explore the change being asked of them and to choose whether to adopt it."

While that last point is the one that surprised me the most and has stuck with me — that you need to give people space to come to terms with the desired change and to decide to either go along or to leave the organization — let's go through the five points in order:

Common and compelling purpose:  Price and Toye write that successful change programs "communicate the why first." And it seems to me that there are lots of powerful whys in insurance for innovation, in general, and generative AI, in particular. Those whys include serving customers faster, more effectively and more humanely in their hour of need; helping them reduce the risks they face with their lives, homes, autos and other assets; and making the insurer more efficient, allowing for lower premiums and a narrowing of the production gap. 

Whatever the why, it needs to be compelling to the employee and needs to be consistent, because the next step is to communicate it relentlessly. Price and Toye write that "change agents typically under communicate their vision by at least a factor of 10." Peter Drucker, the legendary management consultant and author, once told me in an interview that General Electric's longtime CEO Jack Welch would pick one goal for the company and communicate it at every opportunity for five years. Then he'd pick a new goal and pound on that for five years.

Price and Toye said managers also need to allow time for dialogue with employees about what's changing to encourage co-creation rather than just defining and imposing change from on high. "Lasting change occurs through insight, not instruction," they write.

—Operating model: There may be structural changes needed. Changes in processes will certainly have to occur. And there will need to be metrics and rewards. 

That's pretty standard stuff, but care will certainly need to be taken on all three of those points. I can imagine metrics being especially tricky because some of the improvements allowed by generative AI are pretty squishy. 

—Capabilities: What they say here is also pretty standard: Make sure you have all the capabilities you need, match them to the right opportunities and either find or train people to fill gaps as quickly as possible.

I think insurers already do a pretty good job of matching a submission with the right underwriter and a claim with the right representative, but I imagine training will have to be different with generative AI. Training can't be one-and-done because the capabilities of the technology are improving so fast. Training will have to be continual.

—Role modeling: The authors stress the need for clear articulation of what you want your employees to do differently — "You can't expect your people to be mind readers." Then, they say, leaders have to model the new behavior. If you want your people to be gung-ho about generative AI, then you'd better let them see you using it personally. 

—Space: As I said, this is the bit of advice I found most surprising. Every company whose change programs I ever wrote about were clearly thinking of them as top-down, but Price and Toye note, "People have the freedom to choose to not engage, and they exercise that freedom only too often" because of fear of the unknown and the potential for loss. 

Their recommendation: "Change management, like training, has traditionally been seen as a push model and needs to become a pull model, where employees draw what they need rather than having it imposed on them.... If you enable them to experiment with what the change could look like, feel like and be like, then their level of comfort with a new reality is more likely to rise. And if you allow them to choose to change, then they are more likely to adopt change with greater conviction and energy than if you don't." 

What if people choose not to change? Then they've chosen to part ways with you, Price and Toye write, and you need to be disciplined about moving them out as quickly as possible. That may feel odd at a time when the insurance industry is focused on its talent gap, but they argue that you lose more by having lots of people who aren't invested in the changes you need to implement.

Price and Toye certainly weren't writing about generative AI. It didn't exist eight years ago. But I think their ideas are still worth keeping in mind as we implement that technology, in particular, and continue to innovate, in general. 

If even Apple can't count on getting people to jump to a cool new technology, what chance do the rest of us have?