An angel pitch is now being read twice by the time I respond to it. Once by me, once by my AI. I am not the only investor doing this, and I will not be in the breakaway pack for long.
Founders who think this changes nothing should think again. The meeting that emerges from this convergence will be different from the meeting that came before it, and the implications cut in both directions.
A power user, again
I expect to see numerous AI-assisted pitches over the next few years. They are coming whether or not I want them, and the only honest way to evaluate them is to be a power user myself. You cannot assess a category you do not work in. You can only nod and hope.
This isn’t a new posture for me. When PCs first arrived in the workplace in the late 1980s, I became a power user of Lotus 1-2-3. Not because I was on the bleeding edge, I wasn’t. I was working in finance, where the tool had already arrived; what changed was that I made it part of how I worked, not an occasional convenience. Spreadsheets changed what was possible in financial analysis, and the people who absorbed them deeply had a different working life from the people who didn’t. The disposition has stuck with me though the tools have changed. As I get older I am more deliberate about avoiding the bleeding edge. My patience for half-working products has thinned but I still want to be a power user of whatever the workplace tool of the moment is.
Right now that is AI, so I am using it.
What it looks like
I use AI operationally to run MooCoo Ventures, usually for drafting, summarising, structuring, and other unglamorous work that an angel investor in a small operation does too much of. More recently I have started using it to assess pitches. I have fed the Brisbane Angels assessment framework and Seraf’s due diligence content into a RAG, and I tune it with use. When a pitch comes in, I read it myself first, then I run it through the RAG, then I compare.
Today the AI is a second check on my thinking. It’s clear that it will not be a second check for long. That’s fine. The interesting question is not whether the AI is good. The interesting question is what a second check is for.
Differently blind
It would be easy to assume a second check is a redundant check, doing the same job twice in the hope of catching an error. That isn’t the right way to think about it. As a mathematician I prefer the more honest description: the AI is not smarter than I am, but it is differently blind. It does not see the things I see. It does see the things I miss.
My blind spots have a shape. I have priors I have explained away. I have categories I have fewer reps in. I have questions I would have asked on a better day. The AI does not share those blind spots, because it does not share my history. It has its own blind spots, like being too easily impressed by structure, lacking the room sense, or not knowing which founders to take seriously despite a weak pitch. Those are also shaped, and shaped differently.
The value of a second check is coverage of a different blind spot. Two views, two failure modes, with very little overlap.
For deal selection, this matters. The framework I described in If I can’t articulate why it will fail is a structured search for the one or two unsure areas that decide a deal. If those unsure areas live in my blind spot, the framework misses them. The AI’s blind spot is shaped differently. Between us, the coverage gets meaningfully better.
The other side of the table
Investors are not the only ones doing this. Any founder who is not running their pitch through equivalent tools before they walk into an investor meeting is going to be the only person in the room who hasn’t.
This is what forewarned is forearmed means in practice. The pitch that survives the new workflow is the one that has already absorbed the structured critique. The unsure areas are known to the founder before they get to the investor. The narrative has been pressure-tested against the same frameworks the investor is going to apply.
I make up my mind on the first pitch and rarely reverse a pass. I’m being honest about a process where attention is the binding constraint. It is not my job to fix the company. Founders need to put their best foot forward, and the AI-assisted first read makes that even less forgiving. A pitch that wouldn’t have survived a careful human review won’t survive a careful AI-assisted human review either. The bar has not moved, but the speed at which it gets applied has.
Cycle time
When both sides have done the structured work, something useful happens to the conversation.
Today, the investor process is sequential: first read, then follow-up questions, more reading, calls, deeper diligence. The unsure areas surface one at a time, partly because each new piece of information shifts the picture. When both sides arrive at the meeting already knowing where the unsure areas live, the early steps collapse because both have run the pitch through structured assessments. Accordingly, the meeting can move directly to resolution.
This is a considerable efficiency gain. The traditional cycle from first pitch to investment can take months, and a significant fraction of that time is spent surfacing issues that could have been identified in the first hour by either side. Compressing that work changes the rhythm of the relationship. Investors waste less attention and founders waste less runway.
Not unique to angel investing
It is worth saying that this convergence is not a quirk of one investor’s practice. The same pattern is playing out across law, M&A, consulting, and any professional workflow built on structured assessment of complex documents. Angel investing happens to be well suited to it as the documents are small, the frameworks are explicit, and the assessment criteria are reasonably clear. The underlying dynamic is widespread, and the next decade of professional work is going to look a lot like this on both sides of most tables.
What I am describing in angel investing is the early edge of a much broader shift.
What survives
If the AI compresses the discovery work, what is left for the humans?
The most valuable part of the job. The bit I have always found most rewarding.
I often see how a pitch might work. A different angle on the market, a sharper articulation of the bet, a reframing of the competitive landscape, a path through ‘the unsure’ that the founder hasn’t quite seen yet. When that happens, I share the thinking. But I need the founder to digest it and pitch it back to me. The risk is that the founder says whatever they think will get the investment, and the thinking never gets internalised. The reward, when it works, is that the founder hears their own voice articulate the new angle. If they say it, they might believe it. If they believe it, they might do it. If they do it and it works, I have added immense value to the company. That is one of the most rewarding parts of angel investing, and it is not going to be replaced by any AI tool I have seen.
The AI does the structured work. It surfaces the unsure, runs the framework, and makes the second-check coverage better. The human extends the resolution. The mode that survives is the conversation where the investor and the founder build the path together.
What’s changing, and what isn’t
The workflow is converging on both sides. The structured assessment work is going to happen faster, more thoroughly, and earlier in the process than it used to. The cycle from first pitch to first cheque will compress.
The test itself hasn’t moved, the question is still whether I can articulate why a company will fail, against an unchanged 10/90 base rate. The negative test remains the only honest one.
What is moving is who runs the test first, and how completely. The investor used to be the only one with the framework. Now the framework is available to everyone and the founder who has already run their own pitch through it is the founder I most want to meet.
Richard Moore is co-founder of MooCoo Ventures, an angel syndicate that co-invests alongside Brisbane Angels, one of Australia’s most active angel groups. He has made over eighty personal angel investments since 2013.
moocoo.vc