📚 Eyestrain? Let AI Take Over!

Remember the world’s first autonomous racing league we mentioned last week? South African Juandre van Eeden was there, so you can see what it was like with SA commentary. A surprising amount of spin-outs and crashes for AIs, though, ultimately about as exciting as Max Verstappen winning the F1 every weekend…

In this Open Letter:

• The trend: Possibly the most promising AI use-case.

• Instant overseas transfers, dark side of the moon & more signs of a VC boom.

• Funds check: How to know if your startup is VC ”backable”.

• The hardest part of dealing with freelancers: The results are in.

• Want some free stuff? Share The Open Letter.

TRENDING NOW

Out with the Mundande

The fastest reader on earth will soon make your life much easier

If you’ve ever had to read a substantial number of pages (we’re talking hundreds), you’ll know what a painstaking, time-consuming process this is.

It’s lekker to do if you’re an avid reader, and enjoy immersing yourself in a book with a glass of wine or cup of coffee – but not so much if you need to retain the information, interpret it in a meaningful way, or pick up inaccuracies or potential issues.

Just get the Intern to do it

Think clerks in the legal profession or junior journalists. Often paid cents on the rand to what is being charged to clients and working crazy-long hours. It works because it solves two problems:

• You have inexperienced graduates who have to go through the graft to learn and are happy to do so at a low salary. 

• On the other hand, they’re solving major pains for firms in doing a lot of the reading, research, and scouring work that takes a substantial amount of time.

But if there’s something really good at reading a lot, processing data and connecting dots, all at a fraction of the cost, it’s AI.

There’s even a governance play…

It’s not only the clerks' job getting sped up. Being able to process and prompt large reports is something many would find useful.

For municipalities, for instance, a hundreds-of-pages-long audit report is par for the course, and it's not unlikely that those who are up to no good are using the hassle and cost of processing this information to their advantage. 

Heck, consider the transcript of something like the Zondo Commission – over 160’000 pages. Start reading it today and you will be done just before Christmas (236 days of non-stop reading – without sleep).

Let’s face it… no one will read it.

However, there are many people who would love to pull data from such transcripts and reports in an unstructured way (be able to ask questions about it). 

Think citizens of a municipality, journalists, political parties and even you and I.

It’s already happening

Local deep learning startup, Deep Learning Café saw this problem pop up time and again – from working with legal practitioners to helping journalists process the likes of the Zondo report. There’s reason to believe that AI can drastically speed up the process of interpreting large data sets or reports.

That’s when they created DocInsights. It utilises best-practice deep learning techniques coupled with large language models to enable large-scale processing and interpretation of reports or datasets. 

And they have already uncovered some sweet use cases and clients :

• Legal: It's particularly useful in caselaw where the AI-assisted search function helps drill down into the facts that matter from a pool of case documents that would drown a law firm.

• Insurance: Their chronology tool helps plot lifecycles in how data has evolved over time, allowing prompting into customer data sets and identifying irregularities.

• Research: It’s getting used by investigative journalists to uncover stories in weeks that typically would have taken months. How? Well, it can process document information with a high degree of accuracy. You can upload bank statements and invoices and it can find connections between companies' flow of funds to show who paid for what, etc. And in SA’s current landscape, you can imagine just how useful this could be.

When we take away this task from lawyers and journalists in training, the immediate benefit is cost, time, efficiency and accuracy. But this poses a new challenge, how will we train future journalists and lawyers? An opportunity in itself and we’re watching this space.

IN SHORT

⚡️Instant International Money. Money transfer platform Mukuru has announced its instant transfer product that can move money back and forth between SA and the UK, and from SA to any EU country instantly.

🛗 Mobile Career Boost. Telecomm MTN has just launched its Skillsbox portal to help job-seekers uplift their careers with resources available on its web and mobile platforms.

🌚 Shy Moon Mission. This week China will send the backup spacecraft from its 2020 Chang’e mission on a 53-day round trip mission to the hidden side of the moon to collect soil and rocks from the lunar surface.

👀 Crypto Watch. SA’s Financial Sector Conduct Authority (FSCA) has approved 75 crypto asset service providers from the 374 applications it’s received. Both Binance and Yellow Card are notably missing from this list.

🤷‍♂️ We’ve Been Saying. Literally a week after we said to prepare for a VC boom, Tyme Bank investor Norskenn22 said African tech companies and startups are becoming TOO attractive for international VCs to ignore.

👩‍🏫 Learning Funding. Local EdTech HyperionDev has secured nearly R95 million in funding to expand its local reach, as well as extend into the UK.

🦸‍♀️ Efficient Insurance. There’s another new local insurance player, simply called, well, Simply, who says SA insurance sucks, and it’s coming to eat everyone’s lunch.

BUILDER’S CORNER

Is your Startup Venture Capital “Backable”?

Last week we covered how CatalyzU is equipping the next generation of founders for the inevitable rise in VC in Africa. Well, Co-founder Luke Mostert is a reader of the newsletter and reached out to help us put together this week’s Builder’s Corner. Let’s dive in…

I’ve reviewed over 4’000 startup pitch decks in the last 6 years. The most important question I ask myself each time is . . .

“Does this Founder understand what is required for their startup to be a successful outcome for a VC Fund?”

You see, VCs like to talk about “moon shots” or “fund returners”... the unicorns. These terms aren’t just trivial jargon, they’re what determines a VC’s success.

For example, in 2012, Y Combinator calculated that 75% of its fund’s proceeds came from just 2 of the 280 startups it invested into – 0.7%! 

It’s this phenomenon that Sebastian Mallaby coined as “The Power Law” in his book by the same name.

Opposite to a normal distribution curve, The VC Power Law distribution sees 5% or less of the inputs (in this case startups 🚀) bringing about 95%+ of the outputs ($ returns).

But what does this really mean for startup founders pitching to VCs? 

At a minimum, it means funds would like each startup they invest in to have the potential to return their fund.

Here’s how this plays out at a Pre-Seed Fund like Future Africa:

Let’s assume a VC Fund has a total fund size of $10M. Then every investment must have the potential to return at least 1x the total fund size, thus $10M. And let's assume on each deal the fund invests is $200k in the Pre-Seed round at a $5M post-money valuation.

• This means the $200k investment needs to multiply in value 50x to return the Fund (200k * 50 = $10M) and for them to at least break even. But one also needs to factor in dilution…

• 3 additional fundraising rounds (Seed, Series A, Series B) will each add a dilution of ∼ 20% to existing shareholders. So, instead, the investment’s exit valuation must reach a multiple of 50 * (1.2)^3 = 86x in order to accomplish the “fund returner” goal.

In monetary terms, this means:

• $5M post-money * 86x multiple = $430M exit value (meaning the company needs to list on a stock exchange or be sold privately for that value).

If we extrapolate, a software business with annual recurring revenue is typically valued at 10x its annual revenue, hence the startup would need to reach $43M in revenue per year. Thus there should be a path to $43M annual revenue within a VC funds (usually 10-year) lifespan.

For instance: if the startup sells software subscriptions to businesses for $1’000 per annum, is there a clear path to 43k customers?

More granularly, at what intervals (growth rate) does the number of users have to increase each year to achieve this?

Is this rate of growth possible?

And how probable is it? 

And if there is such a path, these 3 core considerations come in: 

• The Founders can pull this off
• The Market can support it
• The Product/Competitive Advantage is big enough to capture enough market share.

Ultimately, though, if a founder can’t demonstrate that their startup is capable of generating a VC-level return, then they will likely struggle to maintain investors' attention for the rest…

If you enjoyed this piece and want to learn more from me and other startup founders who have done it before — such as 2x unicorn founder Iyinoluwa Aboyeji — then apply to our “How to Startup” Fellowship.

Applications close on 14 May 2024, with the program kicking off on 23 May 2024.

YOUR VOICE

We asked about pains in working with freelancers/gig workers, and just finding the right skills is the biggest…

🟩🟩🟩🟩🟩🟩 🔎 Finding someone with the right skill set (34%)

🟨⬜️⬜️⬜️⬜️⬜️ 👎 The quality of work is horrible (8%)

🟨⬜️⬜️⬜️⬜️⬜️ 🤦‍♂️ They’re unprofessional (10%)

🟨🟨⬜️⬜️⬜️⬜️ 🃏 They’re unreliable (15%)

🟨🟨🟨⬜️⬜️⬜️ 💎 The good ones charge too much (21%)

🟨⬜️⬜️⬜️⬜️⬜️ 🤗 No pains, love them! (12%)

Your 2 cents…

Oh, definitely, Fran. And as much as we love it when they offer a special “local” price, we know it’s not sustainable for them. Guess we just need to build our own better local solutions…

FOR THE MEMES

Got startup memes? Send them our way or tag us on socials.