Robert Ambrogi Blog
At the American Bar Association’s Techshow this week in Chicago, the law practice management platform Smokeball will unveil its plan to deploy generative artificial intelligence within its platform across the client lifecycle, from intaking new clients to handling their matters to billing them for the work.
It will launch the first set of these Smokeball AI features in beta in March in the United States, United Kingdom, and Australia, with general availability around the middle of the year, and then subsequent features later this year or the beginning of next.
“Smokeball AI enhances functionality within its existing platform, optimizing workflows to help law firms expedite intake processes, search for legal answers, navigate through matter documents, draft correspondence, and provide tailored descriptions of work directly on invoices,” the company says.
Learn more about Smokeball in the LawNext Legal Technology Directory.
Exclusive Preview
During an exclusive preview of the new AI features for LawSites last week, Smokeball CEO Hunter Steele said that Smokeball’s operating philosophy regarding AI is that it won’t replace small law practitioners, but that those who leverage it for efficiency will eclipse those who do not.
Also driving Smokeball’s AI development is its belief that it is important to integrate AI into the tools lawyers already use, rather than require them to learn new tools, and to ensure that it is delivered with the highest security standards and protocols, he said.
“That’s how we think about AI — how do we efficiently put this into their work processes, rather than putting a whole other tool in another screen or another browser to take them away from what they’re doing,” Steele said. “So we’re engaging with our client lifecycle, we’re engaging with our seamlessly integrated experience.”
The first set of features, called Smokeball Matter.ai, launching in beta in March, will focus on helping lawyers get work done for their clients and manage their matters. It will consist of three components:
- ChatGPT interface. A built-in interface to ChatGPT to generate answers and content from within Smokeball, safely and securely. “We think it’s quite powerful to bring the power of ChatGPT or OpenAI to our users, but in the safe, secure, connected environment they’re used to today,” Steele told me. The interface appears directly within a matter, and no data is fed to the ChatGPT model or the open cloud, Steele said.
- Legal answers. Smokeball will enable users to get authoritative answers to legal questions by using generative AI together with expertise from third-party partners such as legal publishers, legal research companies, and tech providers. Although Steele is not ready to identify those partners, he said, they will include established AI partners as well as newer startups focused on specific legal verticals, such as personal injury. “We’re going to integrate with those in a strategic way to make it easy for you to access them.”
- Matter and document answers. In what Steele believes will be the most important of the new AI features, a generative AI interface will, as he put it, “utilize the power of OpenAI within their matters.” It will allow users to quickly obtain information about a matter, summarize a matter, create content or draft correspondence. “They don’t have to go somewhere else, they don’t have to upload any files, they don’t have to point anything anywhere. It’s there ready for them to point and shoot and experiment with.”
Two other features will also launch in beta in March. They were not included in the demonstration I saw, but they are:
- Smokeball Intake.ai. This feature will use AI to transform firms’ existing PDF or paper intake forms into digital forms that can then be completed more quickly and easily. Although Smokeball already has “powerful” lead intake and matter intake functionality, Steele said, many customers still have PDF or Word intake forms. “So we’re working with AI to automatically take those Word documents or PDFs and automatically create those intake forms for them.”
- Smokeball AutoTime.ai. Smokeball already has the ability to automatically capture time, including what work was done and for how long. In fact, over 50% of time entries in Smokeball are created automatically, Steele said. But this new AutoTime.ai will enhance that feature by automatically creating detailed descriptions of work, so timesheets will show not only time and activity, but a specific and tailored description of what was done during that time. This will reduce the need for lawyers to type their own descriptions, and could eventually increase the number of time entries being created automatically from 50% to 80% or more, Steele believes.
Farther down the road — possibly late this year or early next year — are two additional AI features:
- Smokeball FirstDraft.ai, to draft legal documents and clauses.
- Smokeball Insights.ai, to find staff, client and matter insights.
Seeing It In Action
Demonstrating the matter and documents AI, Steele said its power is based on the extensiveness of the information customers store in the platform. This includes all of their matter information, virtually all of their documents and emails, and full financial, billing and accounting information.
This new feature will apply AI against those three categories of information — matters, documents and emails, and financials. When the feature rolls out, users will see a new tab, Matter AI. This is where the user can enter questions, either directly to ChatGPT with the “Ask OpenAI” option or against the client matter with the “Ask the Matter” option.
With the Ask the Matter option, a user might start by asking a general legal question related to the matter, for which the answer will be provided through the partners described above, or they could ask questions specific to the matter.
In the demonstration, Steele asked a question about non-financial contributions in the context of a divorce. The chat returned an answer describing what that means and providing examples. (A disclaimer on the screen warns that AI can make mistakes and users should check important information.)
Steele then went on to ask questions specific to the matter, starting with a quick question asking for the client’s contact details, which it quickly returned. He then took it a bit further and asked it to summarize the parties in the matter. It quickly showed their names, the role in the proceedings, their contact information, and their representatives.
Steele next asked it to summarize the wife’s non-financial contributions to the marriage. Drawing from the documents in the case, it showed a list of the wife’s non-financial contributions, organized under headings such as “Contributions As A Homemaker.” It also pointed to the document on which it based its answer — in this case, the wife’s affidavit, enabling the user to quickly preview the document and verify the information.
From there, a user could choose to narrow the queries to the specific document. By clicking “Ask This File,” the chat’s answers narrow from the the entire matter to just the document at hand, providing document-level context for the answers.
Based on that, the lawyer may decide the husband needs to beef up his own affidavit with regard to non-financial contributions. Using the same chat interface, the lawyer can tell the AI to draft an email to the client explaining the importance of including additional non-financial contributions and providing examples for him to think about. In just a few moments, the AI drafts the email. Toolbar options let the lawyer drop that text into an email, letterhead, a memo, or elsewhere. In his demo, Steele chose to create an email, and Smokeball opened Outlook, with the draft text added, and the wife’s affidavit automatically included as an attachment.
If the lawyer wanted to change the tone of the email, to make it more or less formal, the AI can do that as well.
Another way to access the AI is when viewing the list of documents in a matter. Right click on any document and there is the option to chat with that document. The user could ask questions about the document or instruct the AI to redraft it in another form, such as from a formal letter to an informal email.
One final application Steele demonstrated was using the AI to create a summary of activity. In his example, he asked it to draft an email to a client summarizing what was done on the client’s matter during the month of January. It reviews documents, calendar events, timesheets and other information and creates a narrative summary of the work.
“So we’re utilizing all of that matter context, all of those documents, everything else, and we’re making it accessible within Smokeball itself and hooking into Word and Outlook to be able to follow on with that correspondence or email or whatever it is at any one time,” Steele said.
Each individual user’s history of AI chats is saved within Smokeball, but users can delete their histories at any time. Only they and the law firm’s Smokeball administrator can see the individual chat histories.
“We are incredibly excited to support law firms as their trusted advisor in legal technology and have quickly realized the utility that AI is going to provide our clients,” Steele said. “Smokeball is uniquely placed to help firms embrace AI due to the advanced productivity and document generation capabilities, our all-in-one and secure environment.
“Ultimately, we can truly help our clients get more work done faster and more accurately than ever before.”