Apps or Yaps: what unlocks AI adoption?

How do you get best practices to flow across a team? Do you need an app for that – or a yap for that?

Researchers from Harvard Business School led a field experiment inside a US sales company to answer that exact question (worded slightly differently to be fair) – and found that the answer is indeed: yaps.

Here's the experiment:

653 salespeople were split across four groups for four weeks. Here were the conditions:

1. Money: this group received financial incentives - tied to a partner's output. If your partner did better, you were rewarded.

2. Meetings: this group had structured meetings with a randomly assigned colleague.

3. Money + Meetings: this group got both.

4. Control: this group got nothing new at all. Poor souls.

So, what happened? Here are the results:

1.      Money

While the money was still being paid, the financial incentives group saw a 13% increase in revenue per call. Money talked – temporarily. When the incentives stopped flowing – so did the knowledge.

2.      Meetings

The structured meetings group, saw a 24% increase in the same period – nearly double. And, this group was still showing gains of 15% or more 20 weeks later.

3.      Money + Meetings

The third condition - where money was added on top of the meetings made no additional difference. The meetings did the work, and the bonus was – well a bonus.

Why the structured meeting worked

What did these magical money-making meetings entail exactly? Absolutely nothing spectacular. At the start of the week, each pair reflected on their most challenging call the previous week, and what they’d do differently. Then at the end of the week, they checked-in over lunch.

When the researchers surveyed the salespeople about the process, they found that most of them already knew who the star performers were and wanted to learn from them. They just didn’t feel they could ask for help. One interviewee described the block as the "intimidation factor". The formal process and the worksheet remove the social cost of asking a colleague for their insight on your challenges.

There is something called permission architecture in action here: the researchers removed the social friction of asking for help (which to be fair, always feels awkward) – by building permission to ask into the process.

How can you borrow the approach for AI Enablement?

Most organisations wrestling with AI adoption are facing exactly the same problem this study describes – best practices are trapped in the minds of your early adopters.

You know, and I know -- there are already people who have figured out how to use AI in ways that change how they work – whether they’ve talked about it or not.

They are getting hours back, producing better first drafts, shortening research loops, doing things their colleagues have not thought to try yet.

Their colleagues might vaguely know that the early adopters are cracking things they really ought to - they just cannot bring themselves to ask, because asking feels like admitting they are behind. Nobody wants to be the person with the Hotmail account, who just does not ‘get AI’ - yet.

How to activate AI best practice sharing:

Step one, naturally, is to come up with a cool sounding name – The AI Exchange?

Step two, pair people up across the team for four weeks (a new partner each week), and give people two questions to work through together: what did you try with AI last week that actually worked, and what have you tried that did not? Encourage them to meet at the start of the week, and check-in again at the end informally.

The organisations that will gain most from AI aren’t necessarily the ones with the biggest budgets and newest apps.

They are the ones that make it easy for knowledge to move between the people who have already figured out how to derive value, and the people who are only using ChatGPT to check if that rash is serious or not.

In summary – when it comes to AI adoption – don’t just think about the apps, also think about the yaps that will unlock their best practice use across the team.

If you’re looking for support kick starting behaviour change – let's yap.

Source: Sandvik, J.J., Saouma, R.E., Seegert, N.T., & Stanton, C.T. (2020). Workplace Knowledge Flows. The Quarterly Journal of Economics, 135(3), 1635-1680.