Model ML co-founder Chaz Englander says agentic AI should automate any task above 60%, with single tasks near 100% automated within 12 to 24 months.
Englander says 2025 was the productivity year and 2026 should be when AI systems start delivering insight that was not possible before.
Englander says adoption at investment firms is a cultural and structural change rather than a technology problem, and that top-down CEO or CIO buy-in separates the best early adopters.
AI Agents Move From Chatbots Into the Investment Firm's Core Workflows
Agentic AI is moving from chat tools into full workflow automation at investment firms, and Model ML is betting that finance will follow the path legal AI has already taken. Chat interfaces cannot handle complex work, so firms need automation that is measurable and embedded in their processes. Gains depend less on the technology than on leadership buy-in and on capturing more data.
Takeaways
Workflow automation beats chat interfaces for complex finance work. Englander says ChatGPT-style chat hits a ceiling on tasks like a 200 page presentation. Firms want full workflow automation, which is easier to quantify and procure. Model ML targets tasks above 60% automation, and a single task could near 100% within 12 to 24 months.
2026 moves from productivity to new insight. Englander calls 2025 the productivity year and expects AI to deliver previously impossible insight in 2026. A European middle market private equity client has automated about 80% of its investment committee memos, including AI-written opinion pages on management teams.
Adoption is cultural and structural, not technical. The best firms have top-down buy-in from the CEO or CIO. They reallocate efficiency gains and designate roughly 10% of staff as AI experts. Englander urges firms to capture more data, especially calls and meetings, which he says hold about 70% of decision information.
Legal AI shows finance's path, a year behind. Englander says 80 to 90% of the top 100 law firms use Legora or Harvey, and finance trails by about twelve months. He says winners design the whole system, from agents and finance context to interface, and embed forward deployed engineers in customer offices.
Model ML claims speed and capital efficiency. Englander says Model ML built a better product in a quarter of the time with a tenth of the capital of its nearest competitors. It raised about $100m after relaunching, ships fixes within half an hour, and grew from 20 to about 100 people.
Sales start with proof and large seat counts. Model ML runs a free two to four week discovery phase and pursues only workflows with at least 60% efficiency gain. Deals usually cover hundreds of seats at $100 to $300 per seat per month, with smaller deals expected soon.
Survival and contrarian checks drive founder success. Both agree a growing business just does not die, meaning it must not run out of cash. Weisburd says his nine billionaire guests compound something and run against the market, and founders should keep rechecking first principles because contrarian ideas look wrong.
Weisburd's media workflow discovery call
Tap a name to go there · tap a line to see why they are linked
On this episode 2
Show 1
Weisburd names video editing, guest mapping, outreach and target list building as the biggest time sinks, and says his team has no real prospecting process.
Englander suggests a workflow that categorises YouTube comments in real time.
Weisburd calls guest booking and relationship graph mapping his two leverage points, since guests are 80% of a podcast's value.
Englander says Model ML plans a product team to map mutual connections for investors and LPs.