Topic
Governance
Approvals, audit, and control for AI that acts.
September 2026: the month agents moved in
The month in AI. Always-on agents went mainstream, frontier models got cheaper and harder to get at the same time, and safety stopped being theoretical.
The AI campaign playbook
In 2026 campaigns used AI to draft. In 2028 the first campaigns will run on it. A department-by-department guide to the operation behind a modern campaign.
The risk isn't the deepfake
Deepfakes get the headlines. The larger risk for a campaign is the AI it deploys itself: answering voters, speaking in the candidate's name, at a scale no one reviews.
The credit manager's AI playbook
Where AI earns its place across a credit manager, from origination and monitoring to CLO operations and the wealth channel, and what has to be true first.
The private equity AI playbook
What a fund should actually do with AI at each stage of an investment, from diligence through exit, and how to build AI EBITDA that survives the next buyer's diligence.
AI across a portfolio, sector by sector
The same technology lands differently in software, services, healthcare, defense, industrials and home services: different work, different constraints, and a different share of the gain the company keeps.
Scaling AI across a portfolio
Most portfolio AI programs can make one company work. Few make the tenth company cheaper than the first. What to share across a portfolio, what each company must keep, and who runs it after the builders leave.
Choosing an AI partner for your portfolio
Lab-owned deployment firms, consultancies, integrators, platform partners or building it yourself. Four questions that separate them, and why model choice matters over a seven-year hold.
Where AI helps across a university
A guide to the work AI genuinely helps with today, area by area, from admissions and advising through research administration and accreditation, and what has to be true for it to help.
No two institutions start in the same place
Higher education is covered as one market with one story. Five kinds of institution, what leadership in each is actually working on, and where AI tends to help first.
Talking is not solving
Voice platforms advertise containment. Containment measures whether a call ended without a human. It says nothing about whether the caller's problem went away.
Why evals matter
The public benchmarks stopped measuring anything useful this year. The instrument you build for your own work is the one that decides everything else.
Meet the work where it happens
Fig is not religious about form factor. Slack or Teams, the mobile app or WhatsApp, desktop or web: the AI fits how your organization operates.
Closing the books with agents
Where agents take work off the close calendar, where people stay in the loop, and what the audit trail has to capture.
Getting the CISO to yes
The five questions security will ask about an AI agent, the four proofs that answer them, and a rollout security can approve.
Sovereign AI: the operating playbook
Sovereignty is a capability, not a location. Five rights, a ladder of infrastructure routes, and four proofs to run before production.