GM. This is Milk Road AI PRO, where we play referee to the AI industry so you have a reliable scorecard to follow & invest with.
In our report last month - The investor’s guide to AI adoption - we built the adoption map for who can take AI from pilot to funded rollout on a 1-3 year timeframe.
Finance is screened as a top-tier sector for one simple reason: It’s basically a paperwork factory that runs on trust.
That’s why AI creates value here.
It doesn’t need a new product to move the needle. It just needs to automate the normal cases, keep humans for exceptions, and let the business process more work with the same team.
That’s where the profit comes from:
Lower cost to serve. Faster cycle times. More output per employee. Stronger operating leverage, without breaking the trust wrapper.
However, there is so much more that AI will impact in the financial industry.
Ultimately, it will be a major driver of blockchain utility and a core determinant of who is going to win the financial “super app” race.
So, here is what we are covering today:
- Service becomes software: Where AI collapses the cost of interaction, paperwork and control, and why this is a margin story.
- Customer distribution wins: How agentic AI turns the interface into intent and routing, and why whoever owns the customer relationship captures the economics.
- Token rails expand: Why software native money fits AI agents better than legacy rails and how stablecoins and tokenization make finance programmable.
- How to invest across the three layers: Mapping emerging winners against the layer they are built for.
Let’s get into it.
Service becomes software
Finance used to be a people business.
Advice, underwriting, support, and compliance delivered by humans, one case at a time.
AI pushes a simple flip: services become software.
Tasks once done by advisors, loan officers and ops teams get encoded into models and agents.
Here’s why this is structural: Digitization moved the front end into apps, but humans still did the “thinking work”.
AI automates that cognitive layer: read, write, explain, triage and sometimes recommend the next step.
So service quality can scale without headcount scaling.
Where does AI adoption show up in practice?
Customer support automation: AI chat and voice support is now normal. It handles routine questions instantly and escalates edge cases to humans.
Sales & relationship management copilots: Copilots help relationship managers and advisors with meeting prep, drafts, and “next best action” prompts, so humans spend more time on judgment and relationships.
Paperwork factory: AI turns PDFs into workflows. Extract, compare, summarize, and draft memos, so teams move faster with humans doing final review.
Credit & treasury support: AI helps scan data, flag early risks, and draft credit work, while treasury tests better forecasting and scenario planning with AI support.
AML, fraud & compliance monitoring: This is the trust layer. AI flags suspicious patterns, cuts noise and speeds detection, while humans stay on decisions and escalation.
Operations + IT automation: Ops moves toward straight-through processing with humans on exceptions, while IT uses AI for coding help and faster incident detection/response.
What are the signals of adoption?
A survey from the Bank of England shows that 75% of financial companies are already using AI.
Bank of America reported back in 2024 that its chatbot “Erica” had 676M interactions with clients and Goldman Sachs reported that a genAI tool completed ~95% of an S-1 draft in minutes (which usually took weeks for teams to prepare).
Those are just examples showing the scale and impact those AI tools can have, with very little adoption efforts and barriers along the way.
So what does this mean?
“Service becomes software” means finance competes more on software velocity and operational execution, not human knowledge, balance sheet, or brand footprint.
The org flips from humans doing the work to humans supervising the work.
AI becomes the execution engine for the “normal” cases; people shift to exceptions, oversight, and trust.
Those developments impact 3 core areas of the financial industry.
First, the operating model gets impacted:
- Work becomes throughput-driven: fewer manual touches, faster cycle times and more cases processed per team.
- The bottlenecks are not staffing, but data quality, approvals and model guardrails.
- Risk management becomes continuous where controls run in the workflow and humans intervene when something looks off.
Second, it reshapes jobs, especially junior entry roles. Less time on drafting, chasing, and copy-paste work.
More time on judgment calls, client relationships, escalation, and model oversight.
A study of Morgan Stanley outlines that globally, around 200,000 investment banking jobs could be at risk over the coming 4 years due to the adoption of AI.
Finally, scaling stops being linear.
If AI handles the baseline volume, adding customers doesn’t require adding people at the same rate, until you hit the edge cases.
Net takeaway
“Service becomes software” is a margin story.
Costs fall first as routine work shifts to AI, then revenue-per-employee rises as teams move faster and touch more clients.
Risk and compliance get cheaper as monitoring becomes continuous and alerts/reports take fewer hours.
Operating leverage improves because growth becomes less tied to headcount until edge cases, data quality and controls become the new constraint.
Customer distribution wins
Most financial products look the same to customers. What feels different is the interface, aka the app.
That “interface layer” is where loyalty and profit tend to concentrate.
“Super apps” are what most people and CEOs refer to in this context. Those are platforms that bundle different services (from various industries) into one integrated offering.
In Asia for instance, super apps like WeChat or Grab bundle transportation, payments, lending, and insurance inside one interface.
In the U.S., Robinhood as an example wins time-on-screen by owning the customer experience and the ever-increasing velocity of new offerings on the platform available to customers.
The chart below shows the strong growth of these platforms over the last several years.
Do we need those platforms in an AI world?
The interfaces today are still a maze of screens and forms. If you want an outcome, you have to click your way there step by step.
Agentic AI flips that model.
You state intent, it plans the steps, executes the workflow, then shows the result and the proof.
So, it’s fair to assume the UI layer disappears next. If the agent can do the work, why keep the dashboard at all?
Uh, Oh… 😧 The rest of this report is exclusive to AI PRO & PRO All Access members!
Already an AI PRO & PRO All Access member? Log in here.
WHAT’S LEFT INSIDE? 👀
- Customer distribution: the financial super app race in an AI agent world.
- Why token rails explode: stablecoins + tokenization go 24/7 as AI agents are not bankable.
- How to invest: biggest winners in each of the 3 layers.
Upgrade your subscription today to unlock access to all of the milky insights above, PLUS:
- More where that came from! Get two AI PRO reports every month, that boil big-picture shifts down to specific investment plays across chips, cloud, energy infrastructure and more.
- Monthly AI AMAs: Your direct line to our lead analysts. Join our monthly "open mic" sessions to grill us on everything from the latest Tesla FSD data to picking the right energy stock. 🎙️
- Access to the PRO community: Where you can chat directly with our analysts and fellow AI PRO members in our private Discord. Ask questions, share ideas, and stay plugged in between reports. 🫂
- The AI portfolio (coming soon): A live look at exactly where we are putting our capital to work across the AI ecosystem. 👜
Already an AI PRO & PRO All Access member? Log in here.
WHAT AI PRO MEMBERS SAID LAST WEEK:
