How can we budget for AI usage and its effect on productivity - and cost management - in the short to medium term?
Which types of headcount will be most affected by AI adoption? What non-headcount expenses will it affect?
Tomasz Tunguz estimates that AI will make Engineering in SaaS start-ups 5X more productive. Engineers are probably the most likely to experiment with AI tools and use it for non-creative, repetitive tasks like routine coding which is typically done by junior developers in larger companies.
Because R&D represents 30-40% of total company headcount, AI adoption is expected to have the greatest impact on R&D headcount (and costs). R&D headcount benchmarks currently are:
- 2023 Median R&D headcount as a % of total company headcount for early stage SaaS companies with revenues $1M-$10M is 38.21%
- 2023 Median R&D headcount as a % of total company headcount for SaaS companies with revenues between $100M-$1B is 29.9%
Marketing is followed by Engineering with expected productivity gains from content development. But because Marketing doesn’t have as much headcount, Tunguz only saw productivity gains in Marketing of 1.5X. However, SaaS Marketing expense is typically only between 35-50% headcount - so headcount reductions aren’t the critical cost saving.
AI could have a big impact on content development, which is a major expense in many Marketing budgets.
- 2023 Median Total Compensation Expense as a % of Total Marketing Expense for SaaS companies with revenues between $1M-$10M is 39.41%
- 2023 Median Total Compensation Expense as a % of Total Marketing Expense for SaaS companies with revenues between $100M-$1B is 37.38%
These benchmarks are for broad sets of the industry with an n of 50+ SaaS companies with validated data from company finance departments. Specific benchmarks vary by SaaS business model, growth rate and a variety of other factors.
What about AI productivity gains in Finance? When we recently asked 2 CFOs, they were cautiously optimistic: “My take is that we as a society are in the midst of an inflection point where technologies like GPT, AI, and ML are advancing so quickly and have the ability to offer tremendous value, but at the same time introduce a new level of risk (data privacy, security, etc.). This puts the finance suite in a very important position where we need to balance risk and return…For example AP automation with a tool like Glean is moving in the direction of streamlining the accounts payable process from procurement to approvals and payments where finance orgs can potentially be cut in half, freeing precious capital to be invested elsewhere."
By contrast, Oracle says “The finance department has taken the lead in leveraging machine learning and artificial intelligence to deliver real-time insights, inform decision-making…”
Who’s right? A random sample of SaaS CFOs, or Oracle marketing? In 2023, we think that most CFOs will be cautious about implementing AI apps in Finance due to a variety of reasons, not the least of which is privacy or confidentiality. There are many routine tasks in Finance which could be done more quickly and cheaply by AI tools, but Finance executives will have to be selective about which apps do not pose any risk to company data confidentiality.
