<- all tokdocs

Kotter's Dual Operating System for AI: Where the Idea Comes From

By , founder of Mavric and Apso

View on TikTok ->

The dual operating system is John Kotter's model for running a hierarchy and a volunteer network side by side, first published in Harvard Business Review on November 1, 2012. A TikTok from @vanessameyersweden argues that AI-native companies should use it in place of permanent innovation labs. Her description of Kotter's network is accurate, but her claims about lab failure rates and collapsing costs come with no source, and one of the two articles she cites argues for something closer to the lab model she rejects.

What is Kotter's dual operating system?

The HBR store lists Kotter's article as "Accelerate!" (product R1211B), dated November 1, 2012. Its summary describes two systems:

  • the existing hierarchy, which keeps running daily operations
  • a separate agile network, which handles large strategic initiatives

The network draws members from every level and is driven by a committed group of volunteers.

The slide in the video credits "Accelerate, 2014", which is the book. A Harvard Business School Working Knowledge article by Kim Girard, dated April 10, 2014, was written for the book's release that week. The idea is about a year and a half older than the book. The video never gives a year, so this is a gap and not an error.

The transcript renders the name as "John Cotter", but the slide spells it Kotter. The video calls him "a professor at Harvard". The HBS page lists his current title as Konosuke Matsushita Professor of Leadership, Emeritus.

How does the volunteer network work?

The video describes "a network that runs in parallel" where people "who have day-to-day operational jobs can also join the network temporarily". The HBS page supports this:

  • The volunteer army is the fourth of Kotter's eight accelerators.
  • Employees join voluntarily, and Kotter frames it as "you get to" instead of "you have to".
  • Volunteers are typically already working within the hierarchy.
  • Kotter calls the two systems an "inseparable partnership".

The video's description also mentions Edgar Schein's "parallel structures" and says membership is "fluid", with initiatives forming and dissolving. The HBS page does not say that membership dissolves, and I did not find a primary source for the Schein reference. Treat both as the author's framing.

Dual operating system vs. innovation lab

The video's slide sets the two side by side. "The lab" is exploration as a separate unit, captioned "Demo admired. Nothing happens." The "Dual operating system" is exploration through the people who will act on it, with "Same people, both systems."

The speaker says labs fail "because they're too disconnected from the rest of the organization." Kotter's model answers that by drawing explorers from inside the hierarchy, so the people who explore are the same people who run the business.

Is the Ambidextrous Organization the same idea?

The video groups Kotter with "The Ambidextrous Organization", which it dates to the "early 2000s" in HBR. Search results identify that article as Charles A. O'Reilly III and Michael L. Tushman, HBR, April 2004, so the date fits.

The content may not fit. Secondary summaries describe the 2004 article as recommending separate units for exploratory and exploitative work, each with its own processes, structures and cultures, linked at the senior-executive level. I could not retrieve the full text. If those summaries are right, O'Reilly and Tushman's model sits closer to the separate lab the video criticizes than to Kotter's volunteer network.

Do innovation labs really fail that often?

The video says labs "have a notoriously high failure rate" and that "production is cheap, cost is collapsing." Neither claim comes with a figure, study or named source. The slide's "Demo admired. Nothing happens." is an illustration, not data.

The same goes for the reference to working "by effectuation principles versus ... planning principles". Effectuation is a decision-making theory from entrepreneurship research, but the video does not cite it, and I did not check how its authors define it.

Who is behind the video?

The speaker is the founder of Afoot (afoot.net), which she describes as "an agentic platform for human coordination." The video points viewers to her Substack article. The description also quotes a Josh Woodward interview about where ideas come from, but the audio never mentions him and I did not find the interview.

The video presents no data, case study or company that has run this design with AI. The speaker calls it "potentially an option". The "AI native" framing is her addition to Kotter's 2012 model.

Bottom line

Kotter's dual operating system is a real, sourced model from 2012: a hierarchy plus a volunteer network drawn from the same people. The video describes it accurately. Its claims that labs fail at a high rate and that AI makes production cheap are the founder's argument, and the 2004 Ambidextrous Organization article may support separate units rather than her design.

Sources

Published October 27, 2026. Based on a TikTok by @vanessameyersweden. Full video notes (markdown)