AlanGoff

Investor · Accenture VenturesCurrent

Driving venture investment and strategy globally, at the frontier of Physical and Agentic AI, with Accenture's most senior leadership.

• Lead and co-lead venture investments globally across Physical AI, Industry 4.0, Data, and Agentic AI — sourcing, diligencing, and carrying them through Investment Committee.
• Set venture investment strategy with Accenture's C-suite and most senior leadership across industries, translating where the firm is placing capital into where its clients are heading.
• Manage a portfolio of 74 companies — leading founder calls, tracking performance metrics, and delivering quarterly reporting to the Investment Committee.
• Author Accenture's venture investment theses for Physical AI and Agentic AI across manufacturing, SDLC, cybersecurity, and fintech — original work that directs where the firm places capital today.
• Write the investment memos that decisions get made on — founder, market, product and competitive diligence, argued to a recommendation.
• Run the diligence side of corporate acquisitions end-to-end — the analysis and materials a target is judged on.
• Scout fintech startups through Accenture's Fintech Innovation Lab, connecting founders to enterprise stakeholders, VCs, and CVCs.

Strategy Analyst · Accenture Strategy

Drove GenAI transformation and fintech innovation for enterprise and PE-backed clients.

• Accelerated digital transformation for a $10B+ PE-backed client by designing GenAI roadmaps — reducing manual processes by 30% across 15 portfolio companies.
• Engineered a Salesforce CRM integration that lifted opportunity conversion 20% and cut data aggregation time 50%.
• Fast-tracked 12 fintech pilots in the Fintech Innovation Lab, driving 25+ C-suite intros and $5M in follow-on funding commitments.
• Led strategy for a Bronx social-impact tech center, training 50+ candidates with a 90% first-cohort retention rate.

Co-Founder & CEO · Toki Labs

Ran a company end-to-end — product, distribution, margins and the exit — before ever underwriting one.

• Built a direct-to-consumer brand from zero to $100K+ lifetime revenue by month 12, growing 10% MoM.
• Negotiated 20 B2B distribution partnerships, generating $30K+ in revenue across 15 states.
• Launched an e-commerce platform at a 5% conversion rate; owned end-to-end product strategy.
• Executed a strategic exit at a 3× revenue multiple, overseeing transition planning and operational continuity.

ML Researcher · Freshman Research Initiative

Undergraduate research that put me inside the model, not just the market for it.

• Co-authored a peer-reviewed paper applying CycleGAN architecture to ice-sheet radiogram enhancement, accelerating image processing 40%.
• Developed a custom ML architecture that improved radiogram diagnostic accuracy 15%.

XBOW

Wrote the investment thesis that led us to XBOW, then ran the deal end-to-end — sourcing, diligence, IC approval, and execution.

Replit

Authored the agentic-SDLC thesis that pointed us at Replit and built the IC pitch and diligence materials. The allocation itself came through a connection — we were fortunate to get it.

Snorkel

Data infrastructure is upstream of everything agentic — the labelling and quality layer decides what the models on top can actually do.

General Robotics

Underwritten directly against my Physical AI thesis: the bet is on the jump from adaptive automation to genuine autonomy, not on the incremental layer.

DaVinci Commerce

Sits inside my agentic-commerce thesis — the transaction layer changes shape once an agent, not a person, is the one buying.

Profitmind

Built the pitch and carried the diligence to Investment Committee.

Agentic Governance Startup

Governance is the gate agentic systems have to pass before enterprises will let them act. The investment has not been announced, so the name stays off the record here.

Scout

A thesis-driven sourcing engine for early-stage VC. Describe an investment thesis in plain English; it searches X, GitHub, and Hacker News for companies that match, scores them against the thesis, and drafts the memo.

• Every figure traces to something it actually read — where there's no evidence it says so instead of estimating, and that constraint is enforced in code rather than asked for in a prompt.
• 942 startups in the database across 13 runs, 643 scored for thesis fit.
• Runs several theses in parallel with version history, and existence-checks investor handles so fabricated ones get dropped.

Aegis

An autonomous underwriting agent that sells AI companies a price cap on compute: pay a premium, and if H100 spot rates blow past your strike during the window, the agent settles the difference on-chain.

• Built for the Hedera AI Agent Bounty (Week 5, Policy Agent), with human approval required before large payouts.
• Every policy, price reference, and settlement is written to a public audit trail anyone can verify without permission.
• 102 passing tests under CI.

A design studio

My own studio, doing packaging and identity for startups — including my own — and merchandise for San Pacho, a house artist with over a million followers who has played EDC Las Vegas.

• Designing the box a product ships in teaches you what a first customer actually judges, which is rarely the feature list.
• Working to an artist's brand rather than my own taste is the useful part: the brief is someone else's, and it still has to be right.

Japan Survey

A guide to Japan I built for friends planning trips: my favourite temples, views, matcha and restaurants, drawn as a survey map you can pan, filter and plan from.

• 74 places across the Golden Route, Tokyo, Kyoto, Nara, Osaka and Hakone, filterable by temples, nature, views, museums, markets, matcha and restaurants.
• Saved stops number themselves and join into a route you can split across days — an itinerary builder, not a pin list.
• Drawn rather than tiled: the ink-on-washi cartography, the bilingual labelling and the scale bar are all hand-built, which is the whole point of it.

It Pays To Be A Winner

A 1v1 turn-based strategy game I built and shipped to Steam in 2019, while I was still in high school. Developer and publisher on the store page are both me.

• Online player-versus-player, so the hard part was netcode and matchmaking rather than art.
• Self-published: I got it through Steam review and onto a commercial storefront alone, at seventeen.
• Still listed and still purchasable — the earliest thing I made that you can go and buy.

This résumé

The page you're on. An agent that answers in my voice and drives the dossier beside it — built because describing taste in agentic products is weaker than demonstrating it.

• The agent runs a streaming tool loop and moves the panel to whatever it's discussing.
• Content rules are enforced by a check that runs on every build, so confidential material fails the deploy rather than shipping.