Not yet, according to the host and the investors interviewed. There is tremendous experimentation and piloting but little hard data on full transformations outside a few bleeding-edge software investors. KKR's Pete Stavros says AI is helpful but still a long way from transformational, an incremental value creation lever rather than the driver of deal outcomes.

Vista's Robert Smith says less than 1% of enterprise data can be absorbed by large language models. The result is very high usage but very little fully scaled enterprise or agentic work. He adds that culture, not technology, is the most important part of change management CEOs face.

Yes, says Robert Smith. Vista requires every portfolio company to have a working agentic angle. In a recent exit, beyond growing revenue more than three and a half times and EBITDA more than five times, the strategic buyer valued the agentic capacity it could spread across its own business for synergies.

Robert Smith, citing Satya Nadella at Davos, says companies must take the models to the data rather than the data to the model. Bringing models into air-gapped or on-premise environments avoids leeching enterprise value into a model someone else owns, and showing how to do this safely will increase adoption.

HG's David Toms relies on friendly competition. To cut DSOs by five days, HG ranked its roughly 50 similar companies and told each CEO and CFO where they sat. Companies asked how to improve and whom to talk to, and top performers happily shared their methods, so no mandate was needed.

John Maldonado says Advent runs over 80 AI initiatives across 30 North American portfolio companies. The most traction is in process automation, consumer intelligence, core operating workflows, and especially churn identification and management. For some of these it is past experimentation and has high conviction in tools that improve productivity, margins and decisions.

The robot is a non-voting observer trained on all past investment committee memos and members' questions to deal teams. It suggests questions to explore, shows what changed since a memo was presented weeks earlier, and flags inconsistent assumptions across deals such as interest rates and FX. Advent started using it three months ago.

Rob Lucas took CVC's leadership to Singularity University around January 2020 and has factored AI into everything since. He uses it daily, tells investment teams on his Monday call that usage is monitored, and says that at CVC not using AI puts you in the bad books.

With nearly 250 companies, KKR brings in an outsider for an AI diagnostic at each and asks every company to run at least one experiment. It tests dozens of vendors across a grid of companies and applications, rolls out winning matches portfolio-wide, and wants AI engineers in its teams and companies.

David Samber says Apollo thinks about portfolio companies, how it invests and how it works. Teams help with content creation, digital marketing, real-time operations and pricing, with about 15 companies exploring automated customer service. Internally, prompts and agents are entrenched in investing, and AI is a force multiplier, not a substitute.