From idea to investor in ten weeks — a cohort case study

One of our recent Academy students went from "I have an idea" in week one to "I have a term sheet" in week ten. The story is unusual. Most students do not get to a term sheet inside the twelve weeks; most need another two to three months after demo day. But this one did, and the way it happened is instructive even for students who land on a more typical trajectory.

We have been asked, by enough subsequent applicants, to write up what actually happened. We have the student's permission to share the trajectory in detail; their name and the specific company stay anonymised, partly because the round is still closing and partly because the lessons travel better without the celebrity attached.

For the purposes of this essay we will call him M. He was 22, had completed two years of a CS degree, was working part-time at a regional law firm doing weekend admin, and had no prior startup experience. He arrived at the Moonlabs Academy with a half-sentence idea and the kind of restless self-direction that you can see across a Slack channel in the first week.

Week one — the wedge

M's first wedge document was, charitably, a mess. The idea was "AI for law firms", which is roughly the same shape as "AI for the entire economy". Too broad to test, too broad to price, too broad to defend.

In the Tuesday afternoon session of week one we made him do an exercise we use a lot: name the buyer, name the budget line, name the alternative they are using today. M produced a document by Friday that read: "Independent UK conveyancing firms with 3-15 staff, currently paying £1,800-£4,500 a month to a paralegal who manually drafts the initial transfer documents from a property questionnaire. The alternative is hiring a junior — about £30,000 fully loaded."

This is a wedge. The economics are visible. The buyer has a name. The pain has a number. By Friday of week one, M was further than half the founders we know who have raised actual seed rounds.

Week two — the first shipped thing

By Friday of week two M had a deployed scaffold at a live domain. Auth, a Postgres database, a barebones form that took a property questionnaire and ran it through Claude to produce a draft TR1 document. It was rough. The output needed editing. The form was visually awful. The conveyancers we tested it on in week three flinched at the design.

But it worked. A real conveyancer could upload a questionnaire and get back a draft document. That is the version-one bar and M cleared it inside two weeks.

The lesson here, repeated by us roughly weekly to subsequent students: the bar for week two is not "good." The bar is "works." M did not waste a single hour on aesthetics in week two. He spent the week on the workflow, and shipped.

Week three — the five calls

Week three is the most uncomfortable for every student. The week they have to talk to real buyers, on the phone, while we are in the room. M did five conveyancer calls in the first three days, two more in the last two. Seven total, all real.

The five calls changed his wedge. The original assumption was that conveyancers wanted time saved on document drafting. The actual buyer pain, surfaced in calls three and five especially, was liability risk — a paralegal making a typing error in a transfer document costs the firm an indemnity claim. The cost of errors was meaningfully higher than the cost of time in the buyer's head.

This was a sharper wedge. The pricing arithmetic became cleaner: we save you an indemnity claim is a much stronger commercial argument than we save you an hour. M updated the wedge document over the weekend.

Weeks four to six — productisation and first paid commitment

Weeks four and five were focused on building the product around the new wedge. Audit log of every change. Side-by-side comparison view that highlighted what the AI had drafted versus the source questionnaire. A confidence score on each field, computed from the model's structured-extraction output. The whole product reframed around risk reduction, not time saved.

By Friday of week six M had a signed pilot agreement at £1,400 a month from one of the firms he had called in week three. Below the eventual list price, but at a tier that genuinely paid for the running cost of the product with margin. The firm committed for three months as a pilot with the right to extend.

This is the moment the entire class shifted in their seats. M was no longer a student. He was the operator of a £16,800 ARR business as of Friday afternoon of week six. The dynamic in the cohort changed visibly. Other students began approaching him on Monday for advice on their own buyer calls.

Week seven — the second customer, and a difficult choice

Week seven is the mid-course pivot window. M did not pivot. He had two more pilot conversations underway. By Friday he had a verbal commitment on the second, at £1,700 a month, and a third in late-stage discussion.

The harder decision was whether to begin a fundraise this term, or wait. We sat down with him on the Wednesday and laid out the options.

Option one: stay heads-down on customers, take the term as a sales cohort, raise after demo day with five or six paying pilots. Lower risk. Reasonable seed at slightly better terms. Almost certainly enough cash to be the right outcome.

Option two: start the fundraise now, in weeks eight through ten, take advantage of the AI-flavoured property-tech tailwind in the current market, raise at the highest implied valuation the market will offer, and risk the distraction from customer work.

M chose option two, with conditions. He wanted at least three more weeks of customer signal first. We agreed to slow-walk the round opening to week nine to give him the customer time.

This was, in retrospect, the right call. But the reason we tell the story is that the decision could equally well have gone the other way, and that founder would also have ended up successful. There is no single correct answer. The discipline is making the decision deliberately, with the cards visible.

Weeks eight and nine — the cold campaign

Week eight, M wrote a list of 60 plausible investors. Property-tech specialist angels, AI-tooling specialist VCs, UK-focused early-stage funds, and a handful of HNWIs from our network we had warmed up. We reviewed the list together on the Tuesday and cut it to 38. The cuts were mostly funds whose thesis did not actually match the wedge despite having a "PropTech" tag on Crunchbase.

The cold campaign went out Tuesday and Wednesday of week nine. Personalised emails, on the discipline laid out in our fundraising essay — research the investor, lead with their writing, name a specific data point you think will surprise them. 38 emails sent, 11 replies, 6 calls booked.

The calls were in the second half of week nine and into week ten. M ran two a day, with us in the room for the first three, and on his own for the rest. We coached the post-call notes each day.

Week ten — the term sheet

The term sheet arrived on Tuesday of week ten. It was from one of the HNWIs we had introduced in week nine — a former operator who had sold a property business in 2018 and was now writing £200,000 to £800,000 cheques into AI-flavoured PropTech.

The terms were strong: £500,000 SAFE at a £4m post-money cap. Generous given the stage. M had a single paying customer and a second on a verbal. The credibility came from the wedge document, the live product, the cohort training, and the fact that we vouched.

M signed the SAFE in week twelve, two days before demo day. He used demo day to surface the round to the rest of the audience and onboard two additional investors — a £100,000 angel and a small institutional follow-on from a fund that had been on the original list. The final close was £750,000 at the same cap, two weeks after demo day.

What the case study actually shows

The story is unusual. Most students do not get to a term sheet inside twelve weeks. We want to be honest about this so prospective applicants do not arrive expecting it. What is not unusual about the story:

• The wedge sharpened in week three by talking to real buyers. This happens in every cohort.
• The first paid commitment landed in week six. A strong majority of students hit this.
• The deck went through three rewrites between weeks five and ten. Every student's deck does.
• The cold campaign worked. Every student we have taken through the campaign has secured at least one investor call from it; conversion to term sheet varies.
• The HNWI introduction was the decisive lever. This is the part of the offer that is hardest to replicate elsewhere.

What is unusual:

• M's wedge was unusually clean from week two onwards.
• M was unusually fast at executing customer outreach in parallel to product work.
• M was lucky in his match-making with the HNWI who became the lead.

Luck is a real factor. We are not selling certainty. We are selling structure, operator time, and access to a network that materially raises the ceiling on outcomes. M's outcome is the upper envelope of what is possible from twelve weeks. Most students hit something closer to "deployed product, two pilots, active investor pipeline, term sheet within four to six months of demo day." That is the median.

It is, we would argue, a dramatically better median than the alternative the same student would have hit doing the third year of their degree.

What to take away

If you are an applicant reading this:

• The unusual thing about M was not his intelligence; it was his execution speed and his willingness to be wrong in public. Both are coachable.
• The structure of the Academy does not produce magic. It produces a high floor and an unusually high ceiling, with most students landing in between.
• Your trajectory is unlikely to look exactly like M's. That is fine. The trajectory you should aim for is yours, lifted by a year of compressed learning. The numbers we are most proud of are not term sheets in week ten; they are the change in trajectory across all twelve students per cohort.

If you want a trajectory like M's, apply. We will tell you honestly whether the Academy is the right fit.

Names and the specific company in this case study have been anonymised at the founder's request while the round closes. We will publish the named version in a year, with the founder's permission.