Consulting · AI products and new businesses

New businesses built on your company's expertise.

With Sam Dix, my colleague and co-lead at the Cornell Tech Startup Studio, I help companies turn their expertise into businesses they can own. You may have a service that AI could make practical to deliver, or an idea that needs a founding team. We join you in developing the product and finding the customer who will pay for it.

We run each venture skunkworks-style, with a small team working on the opportunity apart from the day-to-day business. The team involves a prospective customer early, while their feedback can still change what we build. We can build it as a product line inside your company or as a separate venture.

01 · Starting with a workshop

Starting with a workshop

You can start with an AI strategy workshop, where your team sees working prototypes built on its own business and picks the ideas worth testing. Most recently, a global advisory firm valued at $2.3 billion started this way, and we now work with its business units to build and test the ideas from its offsite.

If you already see the opportunity, we can start there.

See the workshops
02 · A method from the Startup Studio

A method from the Startup Studio

We both co-lead the Startup Studio at Cornell Tech, where we've helped incubate hundreds of startups. Every team tests its assumptions with users and keeps a shared record of its research and decisions. With that record, coaches can follow each team's thinking and pass what one team learns to the others.

We are now turning our coaching into instructions agents can use, so more teams get the questions a coach would ask in office hours. Inside your company, we use the same approach so small teams can work in parallel and reuse what the others learn.

We expect an idea to change once customers get involved, and we help the team work through those changes.

03 · The work we take on

The work we take on

We can begin with an opportunity you already see or develop one from a workshop.

Define the offer

We help the team specify who would buy the service and what they would use it for. We examine the alternatives that customer has today and what your company could do better. When there are several opportunities, we recommend starting with three to five and comparing what we learn.

Make the first product

We work on the prototype with your team, choosing the technology around the job the customer needs done. Prototypes range from teams of AI agents to small models trained on your company's approved knowledge. Your experts review the result as we build, and we keep the scope narrow enough to learn whether the service is useful before committing to a full product.

Form the founding team

As the business takes shape, we help decide who needs to lead it and what capabilities the team needs. We can bring in founders from the Cornell Tech network where appropriate, including when an opportunity should become a company of its own.

04 · Engagement formats

Engagement formats

We work in one of four ways.

Coaching an executive

We coach a CEO, COO, CMO or practice leader while their team builds and tests an idea. Leaders can learn through the work or hand more of the execution to us.

One business unit

We work with a small team inside one unit as it builds prototypes and tests them with clients, using approved knowledge and tools.

A cohort

We coach several executives or teams at once while each works on its own opportunity, and they share what they learn.

A private equity portfolio

We start with the operating team, run a focused experiment in one portfolio company, and take what works across the portfolio.

05 · Earning the next investment

Earning the next investment

We agree on the evidence for continuing before spending the next round of money. Most of the investment comes after a customer pays for a pilot.

  • 1. DemandA prospective customer confirms that the problem matters enough to pay for a solution. We establish that before writing code.
  • 2. QualityA working prototype runs on approved data. Your expert compares its output with the standard of work a customer expects and identifies the corrections it needs.
  • 3. CommitmentThe customer pays for a pilot, and we observe how they use the service. If they won't make that commitment, we stop work on the idea.
  • 4. EconomicsWe count the full cost of delivery, including human review and support. Your sponsor uses the pilot's results to decide whether the business merits further investment or should stop.
06 · Building within an established company

Building within an established company

The team needs a sponsor with authority to fund the work, and enough time from the people who understand the customer. We agree on those commitments at the start. Your engineering team or production partner takes responsibility for the finished product.

One customer works closely with each team as a design partner. We learn from their existing workflow what the service has to fit into and where it needs to improve the experience.

Teams share useful knowledge and software, with your experts deciding what others can reuse. We maintain a record of experiments and decisions so leadership can follow the work. Customer permissions control who sees what. We run the first tests on public or synthetic data, or on data approved for that purpose.

07 · Fees and co-investment

Fees and co-investment

We can work for a fee or for a lower fee plus equity in each venture. We can also take a fee with an option to co-invest, and we settle the scope and terms before the first investment review.

When we want to build the venture together, we can co-fund it and agree early on how each party will contribute and participate in the business.

08 · Booking

Booking

Send the opportunity you're considering to alberto@escarlate.com, along with what you've learned from customers so far. We can discuss who would join the team and whether it makes sense to build together.

Start a conversation See the workshops