Skip to main content
Age of AI Newsletter

Age of AI Edition #5 — Snooping the Tech Industry, Crunching Invoices, and Reducing Clutter

By February 3, 2022No Comments

Here’s your new edition of the Age of AI Series, where I dissect how companies in various industries are already using ML at the cutting-edge (for tangible business value, not research projects).

After doing this for a while, I’m noticing certain themes.

One of them is that for enterprise AI applications, building the right AI models is less than 20% of the work while building the environment and ecosystem in which the model will run is >50%. However, the former can be more expensive (especially getting good data) and its accuracy determines the fate of everything else.

Another theme is around fundraising for AI startups, but I’ll talk that about more in the next edition. 🙂

In this email, you’ll again learn about 3 cool companies.

  1. AI to keep track of “the word on the street” for you as a CEO: Tech is moving so rapidly that as executives, we’ve got to keep track of what’s happening in our industry all the time — a competitor quietly investing in new technology, a new technology very useful for our business strategy, etc. But this is often done passively, through social media or news. This company builds a “B2B search engine” that uses ML to stay ahead of new developments. (Dennis Poulsen, CEO of Valuer.AI”)
  2. AI that handles all invoices for companies: Your accountant will probably agree with the saying, “reading an invoice is harder than driving a car.” Companies waste an unreasonable amount of productive employee hours on this grunt work. But while general OCR has been around for years, getting it to read diverse invoices is a different ballgame — you need a lot of specialized data. (Carsten Nørrevang, CEO of Paperflow)
  3. Making AI less “cluttered” and easier to use in enterprises: There’s a huge number of AI vendors available for everything from facial recognition to transcription or translation. This company has seen tremendous success in just combining all the various services behind a single platform for developers to use, making it more accurate, convenient and less expensive. (Eugen Gross, CEO of Aiconix)

Below you’ll find quick summaries, but please note that the actual episodes pack wayyy more insights!

1-Minute Summaries (if you’re short on time!)

1. B2B Search Engine Using ML

The context:

For most CEOs, the way you hear about new technological developments in your industry (or tech that’s actively being researched, but which could be useful for you) is through happenstance — we read something in an article or on social media and then dig deeper, to see if it’s something we should know about.

Often, by the time we find out about it, it’s already “old” news and everyone’s chatting about it.

The problem:

Turns out, this is a fairly very passive and reactionary way of staying up to date with innovation in the industry. But there’s also a way to do this actively.

How the solution works:

The key idea is to use AI to scour the internet and see what’s going on in various industries, looking for “signals” of new technological innovation. Then you can create a Spotify-like search and recommendation engine, where it automatically points you towards companies similar to you or relevant to you, and which new technologies they’re adapting.

Or, you can look at NEW technologies which are still in the research or early stages (say a certain new manufacturing technique) and track how it’s being adopted across the landscape.

To do this you have to figure out 2 hard things:

  1. How to find “similar” or “relevant” companies, in terms of their business: crawling/mining business registries, job boards, websites, and using that info to make sense of what a company does, so you can put them in a cluster with other companies.
  2. Mining research literature to find technologies that could be useful for different businesses.

We discuss the technology and design challenges in depth:

Another interesting thing is that Valuer.AI is a public company listed on the Danish stock exchange, so Dennis and I talk about the strategy, benefits and other aspects of that as well!

2. Handling invoices for 10,000+ companies

The context:

The finance department of a company can often cost 2-4% of the revenue. And in most companies, the accounts payable team is bigger than the accounts receivable — this is because handling invoices and receipts from various suppliers is time-consuming.

If you can crunch them quicker, you not only save time and money, but also make life easier for accountants.

The problem:

Invoice formats are almost as diverse as the companies themselves. This means using a generic OCR to scan them is as bad (if not worse) than having none. Training better a OCR for this is very data-intensive.

Paperflow’s solution:

  • If people constantly have to “check” the OCR results, it’s a failed investment. So going beyond human-level accuracy is key. They have like 30 people solely dedicated to labeling invoices!
  • You can’t scale forever with human data labeling, so you also need a feedback loop from the customer to improve data quality.
  • There’s a delicate UX challenge in integrating with a human accounting team: building a tool that fits into the process as a smart assistant, still leaving room for the expert.

What I find really cool is that Paperflow is a really great example of how we can use AI to fix inefficiencies across the entire value chain of today’s businesses:

3. One AI provider to rule them all

The context:

There’s an overwhelming explosion of AI products and APIs these days for so many audio-visual content tasks — be it voice recognition, transcription, translation, facial detection, etc.

This is great news overall. For example, German news channels now live-transcribe Angela Merkel’s speech on TV.

If you’re a company trying to use as much AI in your company as possible, you could either build these yourself, or buy ALL these tools separately, OR…. the most convenient would be to just use one API+tool that combines them all under the hood.

The solution:

Aiconix acts as a single platform for many state-of-the-art ML models and APIs, so that you only have to deal with one company for all your ML needs.

What I really like about their product is that it can intelligently combine the results from different providers (say choosing the best translation of a sentence from among Google, Baidu and DeepL) and give you the most accurate result.

Strategic points:

Being a single platform for many services, the big focus is marketing and sales, so most of our discussion revolved around product management choices, sales process, and other things!

Want this newsletter in your inbox? Sign up here!

P.S. Ethics Policy: I don’t own stock in guests’ companies, nor do I get paid by them in any form for any reason at the time of publishing, unless specifically stated. Episodes are also not intended to be an automatic endorsement of any company or its products and services. These opinions are 100% my own as an independent observer and educator.

Leave a Reply

Designed by

best down free | web phu nu so | toc dep 2017