Justin Woodbridge

Founder. Started Bayes (YC S19, acquired by Airtable), and a bootstrapped fundraising site before that. BK, NYC.

Bundling Services and Software

What opportunities are opened up with LLMs? I don’t think it’s software, I think it’s some kind of more scalable done-for-you service offering.

Why not software? The market is flooded with 5M+ seed round AI startups, trying to create pure-play AI software tools for every industry. AI recruiters, AI SREs, sales people. And at the same time, every incumbent vendor is adding in native AI to their products. I think this will be a blood bath:

  1. Very hard to differentiate
  2. “Real Businesses” are getting tired of being slammed with AI vendor pitches. Anecdotally, every time I talk to someone they mention this. The rise of automated cold outbound is not helping this either. Pitching yourself as “AI software” is a red flag / gets their eyes to glaze over.
  3. The tech isn’t reliable enough to sell as a standalone, reliable product.

The (relatively) under discussed alternative is to solve the problem end-to-end for them. Ie, instead of building an AI recruiter product, can you do the recruiting as a service? Historically, this has been less popular than SaaS because:

  1. You have little operating leverage. It is human intensive and you need to scale headcount linearly as you grow. Lots of operational complexity challenges, smaller margins, and limited scale.
  2. Minimal differentiation between any given agency / staffing / bpo company. It comes down to brand.

I think this is where LLMs actually are quite disruptive and opening up new opportunities.

  1. In theory, headcount can stop scaling linearly, leading to much bigger potential scale and reduced headaches.
  2. And, most compellingly: differentiation through software. What if you framed the offer as a bundle of both the software tool, and the service implementation to drive the outcome to completion, completely DFY?
    1. You could also imagine a world where you can undercut competitors pricing for the software, and make it up on the service + software bundle.
    2. The hard part is having the technical expertise and financing to pull off bootstrapping the software.
      1. LLMs are actually make it a lot easier for technical people to spin up bespoke software a lot faster. Not a “death of software” framing, but the trend will continue.
    3. Example: Dover pivoted from being a pure SaaS recruiting product, to offering the software completely for free, and is now a marketplace for buying contract recruiting help to help drive it.

As you go, and the tech gets more reliable, you can probably automate more and more of the work, the software gets more and more reliable, and you can maybe choose to then sell the software as a more standalone product or reduce headcount even more.

Also, as a macro trend, it seems like remote work is bringing a new market demand/problem-awareness to outsourcing/nearshoring generally, which should open up interest to these kinds of offers.

In IT land (my current area of focus), the help desk is one of the best places to do this IMO. Fixify just launched and is doing it, so there is already some motion.