Sterne Kessler Predicts New Hires After Launching AI Patent Tool

Sterne Kessler lawyers say their new internal artificial intelligence tool will save them time and potentially increase the number of human employees they need.

The firm announced this week that it had launched Patent Assist AI, a new technology to streamline patent specification drafting.

The tool was developed in-house by director Daniel Block and patent agent Justin Wang.

Block and fellow director Michael Specht sat down with Managing IP to discuss how the firm developed the tool and what it means for the practice.

Hiring boost

Specht says he expects the launch to have a positive impact on recruitment.

He adds: “I think it will increase the number of people at the firm because it will drive more business to us, and we will be able to produce higher quality applications within the same budget.”

Block says he started to recognise last year how much AI could influence patent drafting.

“I told every single partner I possibly could about the impact this would have,” he says.

“Ultimately, they were sick and tired of hearing from me.”

His colleagues told him to look at the tools in the market and see if using one would make sense.

The firm ultimately tested nine different vendors that offered drafting tools.

To test the tools, the firm ran the same invention disclosure through each of them and analysed the results.

However, as a control, it also inputted the disclosure into a Large Language Model (LLM) that wasn’t designed with patent drafting in mind – and fed the LLM prompts related to how it should draft the patents.

The firm found, however, that the LLM was doing “quite well”.

“That’s when I realised that maybe it made sense for us to build our own tool,” Block says.

Specht adds that developing the tool has given the firm a better understanding of the challenges clients are facing when assessing AI tools. He notes that the firm had to account for ethical issues and ensure that the data would remain private.

“The firm wanted to make sure we weren’t just providing a technical solution but a solution that worked in the legal environment,” he says.

Block and Specht believe that Patent Assist AI will make the firm more efficient. However, they’re still working on quantifying exactly how much time it will save.

The pair say initial feedback has been that it has saved a substantial amount of time.

Pretty accurate

The directors also walked us through how the tool is used.

The user is asked to input the invention disclosure. Attorneys also draft claims based on their interactions with inventors, which they submit to the tool.

After that, the user uploads diagrams of the invention and clicks “generate”.

The process can take around 20 minutes and when that time is up, the tool has produced a draft, detailed specification.

“What’s incredible about the tool is that it’s pretty accurate,” says Block, adding that users can edit the information if it’s wrong.

Specht notes that it isn’t going to be perfect.

“But then the attorney goes through and cleans it up, and you have a patent application and a quite good one,” he adds.

Once the tool produces the draft specification, practitioners can still use AI to refine it, as the tool also contains a built-in AI assistant.

Users can select any word or paragraph in the specification and tell the tool to add additional embodiments or add more detail.

Time management

Block says by developing the tool, attorneys can focus on providing the value that they’re best at.

“The time that’s spent on describing the figures can now be allocated to adding more value to the patent. To me, that’s the real value of the tool and one of the major motivators for me to do this.”

Specht adds that budgets for drafting electronic patent applications are fairly tight.

“This will allow our attorneys to focus, spend a little more time on developing the claims and figures, and interact with the inventors to make sure we get a really good handle on the invention and then use the AI tool to assist them,” he says.

“Picking the inventor’s brain and understanding the invention – that’s where we as people add a lot of value.”