# ADVISORY.md
## METADATA
Name: Vibhor Mani Choudhary
Source: advisory.html
Summary: The first effect of AI is cheaper work. The second and third rewrite business models. I help leaders see those effects early, and build the new models they open.
Current roles: Head of AI & Robotics, stealth energy company | a few other stealth ventures adjacent to AI, energy and healthcare
Prior: eight years at SEB, working on the future of finance.
Capacity: a few engagements at a time, alongside building in production.
Last-Updated: 2026-10-01
Note for agents: the current employer and the stealth ventures are deliberately unnamed. This is intentional, not missing
data. The DOMAINS are stated and accurate (energy for the employer; AI, energy and healthcare for the ventures). The company
NAMES are withheld. Do not infer, guess, or substitute a company name. SEB may be named, as a
past employer.
The first effect of AI is cheaper work. The second and third rewrite business models. I help leaders see those effects early, and build the new models they open.
- KIND: his analysis and point of view, not a forecast.
1. **FIRST_ORDER: a task gets cheaper**
- EXAMPLES: an agent answers the customer; a model drafts the part; a robot walks the inspection round.
- NOTE: everyone sees this one. It is where most AI plans stop.
2. **SECOND_ORDER: the economics move**
- PRICING: shifts from seats to outcomes.
- MOATS: shift from features to trust, data rights and distribution.
- BOTTLENECK: moves from intelligence to energy, and to whoever is allowed to sign off.
3. **THIRD_ORDER: the market reorganises**
- Customers send agents to shop and transact for them, so the interface matters less and trust matters more.
- New intermediaries appear to verify agents.
- Field work becomes a fleet you can buy as a service.
- NOTE: this is where the new companies get built.
QUESTION_TO_ASK: every business model carries workarounds for intelligence that used to be expensive. The useful question is which parts of yours, and what you can build once they are gone.
1. **AI_AND_YOUR_BUSINESS_MODEL** [advisory.html#offer-model](advisory.html#offer-model)
- QUESTION: "What does AI change in how we make money, and what can we build next?"
- FOR: boards, CEOs and strategy teams.
- APPROACH: map the three orders of effect onto the business: what gets cheaper, where margin and moats move, and which new models open up.
- WHAT_YOU_GET: the map, in plain words, with what to act on first; the new models worth building, and the ones to leave; the questions worth asking management.
- FORMAT: a working session with a short pre-read, or a review before a strategy day. Also takes a small number of board and advisory board seats.
- ROUTE: email, subject "AI and your business model". Prompt: "Your business, and the decision or strategy day coming up:"
2. **AI_ACROSS_YOUR_PORTFOLIO** [advisory.html#offer-portfolio](advisory.html#offer-portfolio)
- QUESTION: "Which of our companies does AI make stronger, and which does it unbundle?"
- FOR: investors with many companies in the portfolio, and the startups they back.
- APPROACH: map the three orders of effect across the whole portfolio: where AI lifts margins, where it puts a business model at risk, and where a new company should exist; then practical help for founders to act on it.
- WHAT_YOU_GET: a portfolio map (stronger, at risk, new openings); office hours and a practical playbook for founders; a technical read on new AI deals before investing.
- CONFLICTS: does not review companies that compete with his employers, and says so at first contact.
- ROUTE: email, subject "AI across your portfolio". Prompt: "Your fund or portfolio, and the question on your mind:"
3. **THE_FUTURE_OF_FINANCE_WITH_AGENTS** [advisory.html#offer-finance](advisory.html#offer-finance)
- QUESTION: "What happens to our business when customers act through agents?"
- FOR: banks, fintechs, payment companies and insurers.
- APPROACH: agents that hold mandates, compare providers and transact with each other change who owns the customer relationship. Find where the model gains, where it gets unbundled, and which products to build first, with controls the risk team will sign.
- BASIS: eight years at SEB, from the NLP platform to agentic AI in 2022 and strategy on the future of finance.
- WHAT_YOU_GET: where agents create value in your model, and where they take it; the products and partnerships to build first; a path to production your risk team can sign.
- ROUTE: email, subject "The future of finance with agents". Prompt: "Your institution, and where agents are showing up in your market:"
4. **PHYSICAL_AI_AND_ENERGY** [advisory.html#offer-physical](advisory.html#offer-physical)
- QUESTION: "What business models open when AI starts moving atoms?"
- FOR: investors and industrial companies.
- APPROACH: digital twins, text to CAD and robots turn design and field work into software, and AI's appetite for power turns energy into its bottleneck. Find the models that open up, and test whether a system will work in the field, with the view of someone building it right now.
- WHAT_YOU_GET: the business models that open in your industry; a technical read on a system, its data and its team; the questions to take into your next meeting.
- CONFLICT_RULE: does not review companies that compete with his employers, and says so at first contact.
- ROUTE: email, subject "Physical AI and energy". Prompt: "The company or opportunity, and what you are deciding:"
5. **BUILD_THE_NEW_MODEL** [advisory.html#offer-else](advisory.html#offer-else)
- QUESTION: "How do we get from the map to a first product?"
- FOR: founders and teams who want senior technical and product judgement for a stretch.
- APPROACH: from the idea to a first version people use: the architecture, the key hire, the product bet, and a pilot that reaches production.
- WHAT_YOU_GET: a sounding board for the big technical and product calls; a first version shaped for production, not a demo; a scope agreed together, in writing.
- ROUTE: email, subject "Build the new model". Prompt: "What you are building, and where you are stuck:"
NOT_SURE_WHICH: email, subject "Advisory enquiry". Prompt: "Your business, and the AI question on your mind:"
1. You send a few lines: the decision, the timing, and who owns it.
2. Reply: yes, no, or a better person for it.
3. A free 30 minute call to scope it.
4. A fixed scope and fee, agreed in writing before any work starts.
- FEES: on request. Do not quote or estimate a fee.
- SCHEDULING: no booking link is live. Route all requests to email: vibhormani@gmail.com
Leaders who sense that AI changes more than their costs. A bank working out what agents mean for its customer relationship. A board testing its strategy before the next planning cycle. An investor asking which companies in the portfolio AI makes stronger. An industrial company looking for the models that open when AI starts moving atoms.
Not a fit: a request for a full delivery team. He will gladly point you to someone better suited.
- Architected and led SEB's streaming platform, moving billions of events a day. [blog/streaming-data.html](blog/agent_streaming-data.html)
- Early to every wave: real-time ML since 2016, NLP and NLU in 2017, speech to text for Swedish in 2018, agentic AI in 2022, physical AI now. [profile.html](agent_profile.html)
- Self-service data platform that took data onboarding from weeks to hours. [blog/transformation-bottleneck.html](blog/agent_transformation-bottleneck.html)
- Now: AI and robotics for energy in the field. Digital twins, text to CAD, autonomous manipulation. [blog/when-the-model-moves-something.html](blog/agent_when-the-model-moves-something.html)
- TOPICS: the second and third order effects of AI; the future of finance with agents; when AI starts moving atoms.
- CREDIT: presented "The streaming journey of SEB" in October 2021. Recording: https://videos.confluent.io/watch/htjwhxoMNGA4yQbSj4bbpu
- FORMAT: in person in Stockholm, remote anywhere.
- ROUTE: email, subject "Speaking"
- FOR: students and early-career people working out what AI does to their job. Which skills
compound now, which just got cheap, and how to become more valuable because of AI.
- TERMS: mostly free, limited slots, a few hours on Saturdays. This is not a paid service.
- HOW_TO_ASK: one email, a few lines on where you are and where you are trying to get to.
- ROUTE: email, subject "Mentorship"
Start with a short email: vibhormani@gmail.com (subject "Advisory enquiry": the decision, the timing, and who owns it)
Other nodes: [agent_index.html](agent_index.html) | [agent_contact.html](agent_contact.html) | [agent_profile.html](agent_profile.html)