Notes From the Field: Disrupt, Software Fintech & All Things AI
In a past life managing the Asia Pacific region, I traveled frequently (as in Solitaire status on Singapore Airlines, which for many working and residing in Singapore is a badge of honor like George Clooney’s 10M miles in “Up in the Air”). During my travels I used to send out an internal newsletter called “Notes from the Field” so our dispersed teams would know what’s going on with their colleagues across such a vast, diverse area of the world.
Last week I was in San Francisco attending TechCrunch Disrupt. Although SF isn’t Gurgaron or the Pearl River Delta, the spirit of “Notes from the Field” lives on.
TC Disrupt celebrates the “Start up industrial complex” with categories in SaaS, AI, Security Sustainability, Space and Fintech, among the headliners. The show has a strong presence of founders, corporate development, and seed to mid stage investors (VCs). And it boasts the “Battlefield” competition where 200 start ups vie for the Disrupt Cup and $100,000.
A few thoughts to share from the conference:
The Venture Funding Universe
The start-up Industrial complex is alive and mostly well. Pre-Seed and Seed funding, after bottoming out earlier this year, is starting to creep back up, a leading indicator for a potential to return to more “normal” (pre 2021) levels of investing and valuations. It’s worth noting there are over 31,000 GPs (general partners), up from about 150 less than approximately 30 years ago, according to one panelist. A telling indication of the scale/scope of the start up complex.
Almost every contributor - early to late stage investor, founder and industry insider -consistently repeated themes of: demonstrate value to your customers, focus on key metrics, grow responsibly, and conserve cash.
Fintech
Keeping with the theme that software matters as a path forward for Fintech, two fintech companies disguised as SaaS platforms (or vice versa?) made the Final 20 “Start Up Battlefield” pitch contest final 20 out of 200 prospective entrants. Mainstack enables entrepreneurs and creators to present and sell their work effortlessly with global payment options. Mainstack was founded by two Nigerian immigrants here in the US, when they encountered challenges sending and receiving funds for work. MakersHub: A data-enriched accounts payable operation for companies that care about scaling accounts payable smarter; contextualizing all incoming bill information, down to the line item; eliminating manual data entry; and saving hours of time weekly.
Generative AI
I use Generative AI deliberately, as AI incorporates lots of distinctions. Generative AI is as its name implies; it “generates” intelligence from data, using large language models (LLMs) to provide responses to natural language queries. Some key observations:
As much as Generative (Gen) AI will make some jobs obsolete, will lead to bottomless appetite for others, especially in data engineering. Yet these are new skills, so we’ll all be on industry learning curve (Lior Gavish / CTO MonteCarlo).
Data validity, security, privacy and transparency critical path to achieving enterprise potential. Enterprise data complemented with segment specific models will be needed to serve enterprise needs. Great insights from panel session with #Dynamics FL, an infrastructure provider for data security and privacy for NLP/Gen AI environments.
Gen AI apps will access a “constellation” of Large Language Models ( LLMs) serving an ecosystem of smart agents and purpose built prompt platforms. This ecosystem in its infancy. A key build out: more refined, industry or segment specific LLMs for faster and more accurate responses. ( from a panel moderated by Tomasz Tunguz, GP at Theory Ventures)
Gen AI “phase I” use cases and functions are those that can tolerate “80%” reliability. Human oversight, support or contribution to the task will be required to complete the “last mile” (20%). Phase II will start encroaching on the last mile, as data validity and transparency strengthen. Gen AI is super powerful with unstructured data, contextualizing information and automating tasks. Lots of focus and buildouts in developing domain specific models, smart agents, and infrastructure for data privacy, traceabilty.
Lots of personal productivity “agent” solutions, which naturally fit into the “phase I’ category. In fact I am using otter.AI (a conference sponsor) for transcription, reviewing panels, and note taking services.
SaaS
“We're about to see a pretty significant uptick in the acquisition space, primarily because we saw so many companies pre market change, take investment dollars, that only lasts so long.” - TechCrunch panel on Selling a SaaS Start Up.
Observing a lot more activity in the “acqua-hire” mode, where early(ier) stage companies are being acquired for tech talent and IP. For AI acqua-hires, these valuations can be robust.
One other topic was the use of AI to enhance or automate vertical specific SaaS apps. Kraftful uses AI to collect, analyze and provide insights to the millions of data points the product builder community gets from customer feedback and monitoring systems.
Wrapping Up
And not much of a surprise that a company with “AI” in its name won Disrupt's Start Up Battlefield (“BioticsAI”) . BioticsAI has built an AI-based platform that plugs into an ultrasound machine to prevent fetal malformation misdiagnosis, but has promise in other applications related medical specialties. Super cool.