Use the AI Tools Yourself, Then Show Your Colleagues

Episode Summary

The plainest instruction Kenn So gives, then director of corporate development and strategy at Smartsheet, is for an executive’s own desk: use the tools in your own work and show other employees how you do it, because he thinks a colleague watching you work is more persuasive than anything you could announce. He will not name a single tool, saying everyone uses several at once and no one of them supersedes the rest, and getting a feel for the different flavors is what shows you what they can do. He closes on a cadence rather than a principle: try a new one every quarter. Around that habit he wants a structure, an AI committee of champions with executives on it for buy-in. He is also watching a change in how AI companies get bought, where the acquirer hires the key people and licenses the models, and says hiring them away is effectively the acquisition.

Key takeaways

  • The instruction he gives for an executive’s own desk is the plainest thing on the episode: start using the tools in your own work and show other employees how you leverage them. He puts it underneath the committee and calls that layer micro educational campaigns

  • His reason for the micro educational campaigns is a claim about how people learn. He thinks nothing is more powerful than seeing your teammate or your colleague in the same team using it, because you see the value directly

  • His advice on learning is to stop asking which tool and start using several. Everyone has preferences, he says, and everyone uses multiple AI tools at the same time; no single tool kind of supersedes them all; and getting a feel of the different tools and flavors of it is a powerful thing to see for yourself, and a way to be a model for others

  • Asked what an executive early in the journey should do, his first answer is a structure rather than a tool. Create an AI committee, which he describes as a group of people, leaders and champions who see the value of AI and are committed to pushing forward the adoption of AI tools across the company

  • The composition of the committee carries a two-sided reason. He wants a diverse group representing different functions, so the organization can see the different use cases but also all the different risks of using these tools

  • He puts executives on the committee for a stated reason. Have a group of champions and ideally executives there as well to have buy-in, so the push for adoption across the company has backing

  • On where AI is actually in production at his own company, he calls his own answer no surprise. In February 2025 a lot of the value and a lot of the excitement was coming from the software engineers, on Copilot-like tools, which he called the biggest use case of generative AI at the time across all companies

  • On what the freed capacity is for, he answers from his own desk rather than with a policy. A lot of knowledge work is finding and synthesizing information; horizontal AI tools are good at exactly that; and instead of reading ten documents for one data point he can have it in thirty seconds

  • His newsletter subject at the time is a change in the shape of AI acquisitions. Instead of mergers and acquisitions, the structures are being called licensing and acquisitions, or L and A

  • The acquisition mechanism he describes has two halves that happen together. The acquirer hires the key people out of the startup and separately pays licensing fees for the models those people built, which he illustrates with Google hiring the Character AI founders

  • His conclusion about the structure is blunter than the deal terms make it sound. When you hire the people away you are effectively really acquiring the company, in essence

  • For an executive who wants the business-strategy view rather than the adoption view, he names Stratechery by Ben Thompson, because it focuses on the business impact, and then adds what he calls a shameless plug for his own newsletter

  • On board-level risk he says he does not know whether a single resource exists, and offers two things. Run a workshop with the AI committee, and look at the new trend of AI-focused law firms, which he says can be complemented with the usual outside counsel and then synthesized for the board

  • His closing ask is a cadence rather than a principle, and he sizes it deliberately small. Keep trying the tools yourself, not every one of them, but probably a new one every quarter, because a significant new product or user experience arrives about that often

About Kenn So

Kenn So was director of corporate development and strategy at Smartsheet, the work management software company, at the time of this conversation, and is now with OpenAI in corporate development. He also writes Generational, a newsletter on the business strategy and technology of AI that he started to clarify his own thinking and find the signal in the noise, and kept going because people kept reading. He has followed the field since before deep learning was popular. He comes at enterprise AI from the corporate development side, which is where his interest in how AI companies actually get bought comes from, and his argument on this episode is that AI spreads through a company by being used and shown rather than announced.

 

In this episode

00:41 Welcome, and who Kenn So is
02:12 What Generational has been covering lately
02:17 A new acquisition structure for AI startups
02:53 Licensing and acquisitions: hire the people, license the models
03:22 Hire them away and you have effectively acquired the company
03:34 The people, not the technology
04:31 Which departments are using AI in production
04:36 Engineering first, and why that is no surprise
05:23 Where the freed-up capacity goes
05:56 Ten documents, or one data point in thirty seconds
06:18 Guidance for executives early in the journey
06:26 Create an AI committee
06:55 A diverse group, so you see the use cases and the risks
07:18 Champions on the committee, and executives for buy-in
07:43 Micro educational campaigns, and watching a colleague work
08:06 A suggestion from the host: rotating the executive sponsorship
08:46 What resources he recommends for executives
09:52 Learn by using the tools yourself
10:20 Why he will not name a single tool
10:45 Stratechery, and a plug for his own newsletter
11:32 Risk management, and what to bring to the board
11:50 A workshop with the AI committee
12:05 AI-focused law firms as a new kind of resource
12:32 What listeners should take away
12:40 Try a new tool every quarter

In Kenn’s words

“I think it’s to create a- an AI committee”

Kenn So   (06:26)

“using the tools yourself is probably the biggest thing you could do at the micro level”

Kenn So   (09:52)

“getting a feel of the different tools and flavors of it is a powerful thing for yourself to see”

Kenn So   (10:20)

“nothing is more powerful than seeing your teammate or your colleague in the same team using it”

Kenn So   (07:43)

“So when you hire them away, you’re effectively really acquiring the company, in essence”

Kenn So   (03:22)

“probably you just need to try a new one every quarter”

Kenn So   (12:40)

 

Resources

Recommended on air

  • Stratechery: Ben Thompson’s publication, which he names at 10:45 as the resource for an executive who wants the business-impact view rather than the adoption view

  • AI Realized Summit: The event he recommends at 09:13, for curated content and a network of enterprise executives outside the tech hubs

Named on air

  • Character AI: The startup in his worked example at 02:53. Google hired the Character AI founders and key people and licensed the models they had built

  • Copilot-like tools: His description at 04:47 of what the software engineers at his company are using

  • AI-focused law firms: The new trend he points at for board-level risk at 12:05, which he suggests complementing with the usual outside counsel

Ideas and terms discussed

  • The AI committee: A cross-functional group of leaders and champions, with executives on it, responsible for pushing adoption and for seeing the use cases and the risks together

  • Micro educational campaigns: His name for the layer below the committee. Use the tools in your own work and show colleagues how, on the grounds that watching a teammate is more persuasive than being told

  • Licensing and acquisitions: The structure he says is replacing mergers and acquisitions for AI startups: hire the key people, license the models they built, and the company is effectively acquired

  • Rotating the executive sponsor: Raised by Christina Ellwood at 08:06 as something she has heard companies do: each executive takes a turn on the committee and learns by doing it. His build on it at 08:27 is that each one would share how they used the tool and how it changed their team’s workflow

  • A new tool every quarter: The cadence he closes on, sized so that an executive can keep it. Not every tool, but enough to know what a new generation of them can do

 

Frequently Asked Questions

 
 
 
 
 
 
 
 
 
 
 
Previous
Previous

AI Is a Tool for Augmenting People, Not Replacing Them

Next
Next

Synthetic Data Has a Place, but It Is Not the Bias Panacea