Harvey AI isn't the only game in town anymore. A wave of specialised LegalTechs is rising to meet the firms it doesn't serve.
A profound comment was made by CEO Winston Weinberg of Harvey AI. He stated that you must "re-earn your role every six months."1 AI is a fast-moving industry, what worked yesterday will not work tomorrow, and to stay on top, AI must constantly reinvent itself.
The reason I state this is because in the last six months there has been an unbelievable rise in LegalTech AI being created, and Harvey AI is right to be concerned. As spoken about in previous essays, due to Harvey AI being on the expensive end and more suited for bigger law firms, many medium sized specialized law firms are potentially looking for a more specialized type of AI. You will notice a trend based on this below:
Jacob Klionsky wrote a great article on LinkedIn on this topic, and he spoke about ten of the big LegalTechs he has been following.
I just mentioned Jacob's top six above, and if you notice, there is a strong trend going on. Almost all of these LegalTechs have become more focused and specialized, rather than one overall model for a larger enterprise law firm. Whether it is based on patent drafting models or litigation fact graphs, these LegalTechs understand there is no need to compete with Harvey AI, rather, let's become unique instead!
Let me expand on each of these below:
| Company | Focus | Funding |
|---|---|---|
| Sandstone2 | In-house legal triage | $30M Series A |
| Newcode3 | Legal AI harness | $13.5M Series A |
| Stilta4 | Patent research automation | $10.5M seed |
| Fearn5 | Patent drafting models | $5.5M seed |
| Inhouse6 | SBM legal platform | $5M seed |
| Turbo Law7 | Litigation fact graphs | $3.8M pre seed |
What is interesting to me is that this is a very recognisable pattern from other areas of AI. A generalist leader begins by emerging, then a wave of specialists come along, and each does one thing better than the generalist ever could. It happened to banking with fintech, and now it appears to be happening to legal AI too.
It is also important to recognise that these six companies are not the ones capturing most of the money in legal AI. Across a recent twelve month period, the ten largest funding rounds accounted for over seventy percent of all disclosed capital in the sector,8 and those rounds went to the biggest names, not the specialists in this table. Most smaller, specialised LegalTechs are working from a much smaller pool of money, and not every one of these six will necessarily make it big.
At the same time, Harvey and Legora's revenue keeps growing fast even whilst this wave of specialists raises money. Harvey grew from roughly one hundred million dollars in annual revenue to around three hundred million within about a year,9 and Legora recently crossed one hundred million dollars in annual revenue of its own.10 This suggests the market is not simply moving from the generalist to the specialists. It is segmenting. Larger firms wanting one platform for everything still appear to be going to Harvey and Legora, whilst smaller or more specific workflows are increasingly going to the specialists.
Put together, this does not look like specialists beating the generalist. It looks like the legal AI market turning out to be bigger, and more varied, than any single company can serve alone.
For anyone entering the profession, the practical lesson is not to learn one platform inside out. It is to build the habit of evaluating and picking up new tools as they appear, because the tools themselves are clearly not settling on one winner. In a strange way, that is exactly what Weinberg himself was describing all along, just at a scale beyond his own company.