For the past 25 years, application software startups have had a singular focus: increasing company and employee (including developer) productivity. This looked like building software that increased productivity at the employee level, increased collaboration across employees and teams, and/or enabled better oversight and management at the leadership level. More often than not, this software has been priced on a per seat basis, in essence benchmarked against the cost of the headcount itself and increasing that headcount’s productivity.
I think my experience at Notarize is relevant here, not because it is AI but because it offloads a core task (i.e. a mortgage closing) from someone. We "sell the work" by offloading the task (albeit to other people, not to AI) and allow our customers to repurpose employee time to other accretive tasks or to alter their staffing models. Much of our ROI story is rooted in that outcome. The primary challenge though is to actually get customers to adopt above a threshold required to change either their internal operations or staffing models. The beginning of a customer's adoption curve is the worst - they're forced to run bifurcated processes (actually costing more) AND the actual people you aim to offset are often critical in managing that transition. At Notarize, I've cribbed some thoughts from the medical and hospital industries which believes that a 30% adoption rate is required to see the ROI of actual process change, beyond which they can actually adjust the "standard of care" and make that new/better process the new normal. Getting to 30% is really hard and we've obsessed over doing just that, which is to convince our customers to make us the standard of care for mortgage closings, auto sales, you name it. I think it will be especially hard for AI in some of the industries you outline above, particularly legal/compliance/etc. Why? Everyone says LLMs drift, but that will surely be solved. Many of the things you described are considered the practice of law and people will need to adjudicate UPL claims. Fun!! I think the real issue is regulatory headwinds. Specifically, regulators are terrified of algorithms/machine learning/AI systems instituting global systemic bias. Take a property appraisal, which everyone agrees should be digitized. Regulators would rather local consumers connect with local appraisers to disperse and decentralize the bias - an African American consumer may get an African Appraiser and "win" on one and another might get a racist white appraiser and "lose" on another. That is better to them than a one nebulous system making judgements they cannot assess. And the government has no ability, mandate, or agency even to test these models. So how is a large bank who is constantly sued for unfair lending practices able to adopt these systems? If anything, recent advancements are showing any/all of these issues can be solved... but for AI to advance as you outline, it needs to think much more deeply about bias and how to instill the confidence required to take over from the humans who are obviously slow, but easier to regulate and more random in their outcomes.
In a way this has already been done in other industries besides AI - services like DoNotPay or Orkin have taken traditional industries, figured out optimizations at scale / how to automate the "dirty work", and profited in huge ways. The issue is - scaling the sales and implementation (well) is harder than just throwing AI at it. Great read!
I run a startup (Lexoo) where we both sell the legal work (in our case outsourced BAU contract negotiation) and sell software, so can provide some insight on this. For the 'selling the work' we use a team of in-house lawyers who use a lot of our own tech to be efficient. We can definitely charge a lot more there than on the software, as we indeed compete with services companies like law firms as opposed to software prices.
Separately, we license some of our tech to customers who may not want to outsource. Here, we are forced by the market to charge typical per seat pricing.
However, one limitation on the services side, is that a lot of customers (our customers are companies with in-house legal teams) have fundamental reasons why they don't want to outsource. So the market size in terms of # of companies who can buy outsourced legal services for work they typically do themselves, seems lower than the # of companies who consider buying our software.
The other hard bit about selling the 'work', is the expected standards in terms of how 'custom' it is are way higher compared to what clients expect from our software. So we end up having to hire quite experienced lawyers to basically enable the 'final 10% of quality control to happen. So that's a bit of a scaling constraint to selling the work.
In the lens of productivity cost and job-to-be-done:
SaaS is just cheaper expression of productivity compares to your own resources. Naturally you would opt to use them to get the job done.
LLM reduces the cost of productivity by 2-3 orders of magnitude. Startups should focus on finding the job that require high productivity cost and address those with LLM.
How do firms differentiate themselves though if everyone uses this? If I’m a lawyer and, after the gains from a first mover advantage, how do you market yourself? If you’re selling the work then it feels like every law firm becomes a glorified marketing/brand play. Some might argue that’s true today, but I still feel like reputation, experience, etc were the defining characteristics of that specific profession.
Would love to connect with you and the founder of Automate Ventures to discuss further in the New Year. We're launching professional services companies that follow the model you describe.
Great Post / Tech Forecasting - So much going on here that can be extrapolated even further. It will be interesting to see which SaaS apps adopt new tech and evolve and which get passed by new AI based services. Also interested to see how the sales model evovles - will it too use AI? AI selling AI?Traditional SaaS (great productivity enhancer that I use in my company) seems a bit tired - there is less and less differentiation as all claim to do everything.
Love the concept! Wondering which party should be held accountable if errors occur - the startup that created the "work," the end user using it, the AI model itself, someone else? There can be insurance opportunity created around that too.
Great read. Law is also a great domain for this, because costs like this are so often passed onto the end client. The end client is happy because it probably cost less than what a lawyer would have cost. The law firm is happy because they can increase matter flow and focus on adding high-leverage strategic value, which is worth a much higher hourly rate.
This is so smart.
I just keep reading and rereading it.
So smart.
I think my experience at Notarize is relevant here, not because it is AI but because it offloads a core task (i.e. a mortgage closing) from someone. We "sell the work" by offloading the task (albeit to other people, not to AI) and allow our customers to repurpose employee time to other accretive tasks or to alter their staffing models. Much of our ROI story is rooted in that outcome. The primary challenge though is to actually get customers to adopt above a threshold required to change either their internal operations or staffing models. The beginning of a customer's adoption curve is the worst - they're forced to run bifurcated processes (actually costing more) AND the actual people you aim to offset are often critical in managing that transition. At Notarize, I've cribbed some thoughts from the medical and hospital industries which believes that a 30% adoption rate is required to see the ROI of actual process change, beyond which they can actually adjust the "standard of care" and make that new/better process the new normal. Getting to 30% is really hard and we've obsessed over doing just that, which is to convince our customers to make us the standard of care for mortgage closings, auto sales, you name it. I think it will be especially hard for AI in some of the industries you outline above, particularly legal/compliance/etc. Why? Everyone says LLMs drift, but that will surely be solved. Many of the things you described are considered the practice of law and people will need to adjudicate UPL claims. Fun!! I think the real issue is regulatory headwinds. Specifically, regulators are terrified of algorithms/machine learning/AI systems instituting global systemic bias. Take a property appraisal, which everyone agrees should be digitized. Regulators would rather local consumers connect with local appraisers to disperse and decentralize the bias - an African American consumer may get an African Appraiser and "win" on one and another might get a racist white appraiser and "lose" on another. That is better to them than a one nebulous system making judgements they cannot assess. And the government has no ability, mandate, or agency even to test these models. So how is a large bank who is constantly sued for unfair lending practices able to adopt these systems? If anything, recent advancements are showing any/all of these issues can be solved... but for AI to advance as you outline, it needs to think much more deeply about bias and how to instill the confidence required to take over from the humans who are obviously slow, but easier to regulate and more random in their outcomes.
In a way this has already been done in other industries besides AI - services like DoNotPay or Orkin have taken traditional industries, figured out optimizations at scale / how to automate the "dirty work", and profited in huge ways. The issue is - scaling the sales and implementation (well) is harder than just throwing AI at it. Great read!
This is a phenomenal take.
Great read!
Great article!
I run a startup (Lexoo) where we both sell the legal work (in our case outsourced BAU contract negotiation) and sell software, so can provide some insight on this. For the 'selling the work' we use a team of in-house lawyers who use a lot of our own tech to be efficient. We can definitely charge a lot more there than on the software, as we indeed compete with services companies like law firms as opposed to software prices.
Separately, we license some of our tech to customers who may not want to outsource. Here, we are forced by the market to charge typical per seat pricing.
However, one limitation on the services side, is that a lot of customers (our customers are companies with in-house legal teams) have fundamental reasons why they don't want to outsource. So the market size in terms of # of companies who can buy outsourced legal services for work they typically do themselves, seems lower than the # of companies who consider buying our software.
The other hard bit about selling the 'work', is the expected standards in terms of how 'custom' it is are way higher compared to what clients expect from our software. So we end up having to hire quite experienced lawyers to basically enable the 'final 10% of quality control to happen. So that's a bit of a scaling constraint to selling the work.
This is a really great post. I should come back to read this again.
Memo to myself: https://share.glasp.co/kei/?p=Xo7I0Nx15vdr8FdpZLs1
I talked about this before: https://rpgbx.substack.com/p/why-personalized-ai-companies-win
In the lens of productivity cost and job-to-be-done:
SaaS is just cheaper expression of productivity compares to your own resources. Naturally you would opt to use them to get the job done.
LLM reduces the cost of productivity by 2-3 orders of magnitude. Startups should focus on finding the job that require high productivity cost and address those with LLM.
How do firms differentiate themselves though if everyone uses this? If I’m a lawyer and, after the gains from a first mover advantage, how do you market yourself? If you’re selling the work then it feels like every law firm becomes a glorified marketing/brand play. Some might argue that’s true today, but I still feel like reputation, experience, etc were the defining characteristics of that specific profession.
Loved this.
Would love to connect with you and the founder of Automate Ventures to discuss further in the New Year. We're launching professional services companies that follow the model you describe.
Will DM you!
Sadly this is going to eat much of India's lunch, especially the service sector's.
Great Post / Tech Forecasting - So much going on here that can be extrapolated even further. It will be interesting to see which SaaS apps adopt new tech and evolve and which get passed by new AI based services. Also interested to see how the sales model evovles - will it too use AI? AI selling AI?Traditional SaaS (great productivity enhancer that I use in my company) seems a bit tired - there is less and less differentiation as all claim to do everything.
Love the concept! Wondering which party should be held accountable if errors occur - the startup that created the "work," the end user using it, the AI model itself, someone else? There can be insurance opportunity created around that too.
Great read. Law is also a great domain for this, because costs like this are so often passed onto the end client. The end client is happy because it probably cost less than what a lawyer would have cost. The law firm is happy because they can increase matter flow and focus on adding high-leverage strategic value, which is worth a much higher hourly rate.
Thanks for writing this, Sarah. Going to recommend EvenUp to my brother who has a personal injury law firm and I'm sure it'll be helpful.