Undoubtedly begins the agent of he ai + al interview – artificial lawyer

Undoubtedly, which focuses on the review of complex contracts, has launched ‘enhance’, a new agent AI-agent system-joining a growing movement of agents throughout the legal technology market. Plus – See the in -depth interview of Al below with Security ísaksson, chief of artificial intelligence.

First, it is great to see how fast agent systems are spreading. Why? Because single requirements only get you so far in terms of efficiency. An agent that can be personalized in your needs and then get a series of tasks, or a task again and again, adds additional efficiency in your workflow.

So what do you do? The United Kingdom-based company, which now has a global reach, explained that with ‘improvement, each model-or agent-specializes in specific tasks, such as summarizing or designing support. Defefy’s agent system analyzes the task, creates a plan and appoints agents to handle different parts of it. ‘

Plus – and Al should say, this site always likes when companies say this little… .the ‘agents cooperate‘Something is something very nice to be able to use numerous agents working together to complete a complex task. It is a modern version of the Adam Smith Treaty for work specialization. Although in this case it is the specialization of the legal work performed by a coordinated group of his agents.

Plus, this relates to the idea of ​​’energy sharing’ that AL has recently explored, ie. If we truly benefit as much as possible from the legal remedies, then we must give them ‘responsibility’, as it was, to do a task and to complete that task-or most of it, or in other words, to give them ‘agency’ 🙂 ..!

–

Undoubtedly Webinar: Success Tactics – Change Management + AI – June 17

–

It is certainly not all issues to print a key and outside them. The performance of agents should be refined with the contribution of human experts. (And as mentioned yesterday, agent of Microsoft contract builder – this is actually an essential part of the process when doing legal work.)

As it is finally noted: ‘The performance of he (is) is evaluated and improved using the legal data of the Deffining Owner and internal expertise.’

And, ‘lawyers (can) interact with the agent system within Microsoft Word’ – which will be essential, AL would add, to make sure you are getting the results you want.

In general, what you get is ‘an improvement in 40 – 70% speed in the course of work against a traditional manual approach’.

SmallCo -founder and leading strategy officer undoubtedly commented: ‘Lawyers work in high -pressure, high -interest environments, where a minor mistake in formulating a contract can cost a legal firm. Finally, we are committed to building products focused on what will really help our customers risk their daily work. As a company, we have not rushed the first head to build it; On the contrary, we have spent time identifying market segments that will flow the highest value for lawyers and working with them to solve the most important problems. ‘

–

Interview in the depth of Al with Security ísaksson, chief of artificial intelligence in obviously.

You declare that ‘agents cooperate and collect information, passing it up to the end’ and then the performance of the agents of it is evaluated and improved using the legal data of the Defey owner and internal expertise ‘-this sounds very interesting-can you explain how this happens?

Agentdo agent within our ecosystem is designed to address a special essential task for comprehensive legal work. By adding new agents, we constantly expand the system skills, enabling it to manage the increasingly complex legal flows and questions.

Most importantly, these agents have access to a strong group of tools, which include our owner’s products. Communication with these products is orchestrated using MCP (model context protocol), which ensures smooth integration. Definly’s suite, specially developed to address the flow of complex legal work, provides our agents to perform legal tasks from end to end efficiently and accurately. This unique integration between our agents and the advanced Defeely tools is critical – it enables the completion of the flow of comprehensive work that significantly direct the load of a lawyer’s lawyer, something accessible only because of the basic bases set through the preliminary development of our particular product suite.

How would you define the ‘agent’ in this context?

An agent is a modular, autonomous component within our system that is created to perform a specific function or a set of tasks, as part of a larger, orchestrated work flux. Agentdo agent is built for a clearly defined purpose (eg, document understanding, precedent, legal research) and is equipped with its logic, tools and access in context, allowing it to make decisions, process information or interact with data in a goal -oriented manner.

How much can agents personalize and control users?

We have integrated a human-in-Loop process, which means that users can review and adjust the plan generated and certify or refine the results of individual agents. This keeps the flow of work transparent and adaptable while using the strengths of the automation.

Our agents integrate directly with our Vault product, which contains all the precedents and gold standards collected by the legal firm, along with the precedents individually preserved by lawyers in their personal porches. As a result, results such as legal texts designed or specific task solutions directly draw from these adapted precedents, reflecting the historical practices of each user rather than relying on generic responses from a linguistic model with general goals.

Why do you do this and why now?

Recent advances in large language models, sophisticated agent-agent communication and smooth integration of external tools through the model’s context protocol (MCP), all critical elements, are now in line. These technological advances enable the creation of powerful, cooperative multi-agent systems that can solve the inclusion of complex work from bottom to bottom, at a time when businesses are eager for solutions directed by those that exceed traditional chatbots and provide true operational transformation.

Which llm are you using for this?

We use a range of main LLMs, choosing the best model for each task based on its strengths. Some models are better in reasoning, others in summary, or handling long documents. Our system is designed to run under -task – such as legal research, matching precedents or obtaining the context – with the most appropriate model, providing high accuracy, efficiency and control.

How do you ensure accuracy?

To ensure that our system performs reliably and accurately, we have created rigorous evaluation processes and comparison. Using automated assessment pipelines and our owner’s legal data, developed through extensive cooperation with internal legal experts, we constantly test and refine our platform. Our evaluation process controls include:

• Plan Correctness: Assessment If ‘orchestrators’ agents generate a sensitive and effective subsidiary division.
• Road of Execution: Assessment If the task trajectory through the system follows an optimal path.
• Exit accuracy: Performing detailed estimates to verify that the results closely match our curated expert references.

Thank you!

–

Undoubtedly Webinar: Success Tactics – Change Management + AI

On June 17, the lawyer is undoubtedly and artificial partnership to bring you a free webinar to succeed in him and setting legal technology, managing changes in the center of discussion.

Speakers will include:

• Jayanth Poorna – Partner, culture and practice of change, Korn Ferry – as an engineer from qualification, a software developer training and a specialist in culture, leadership and change, Jay brings the unique ability to see technology, operating models and transformation through human experience lenses.
• Molly Taylor – Chief of client success, eventually – who runs the client’s success team in a permanent manner, where it is committed to ensuring that customers achieve measurable values ​​from defecly legal products. Before joining obviously, Molly held roles in DXC Technology, Workshare, Letters and Legl.
• Panel chair – Richard TromansFounder, artificial lawyer.
• Plus – a guest guest… And yes, I don’t know who it is either

To register on the Live Webinar at June 17, at 3pm in the United Kingdom (BST) and 11:00 ET, Please RSVP here.

–

On the webinar we will cover:

1 – why legal technology initiatives fail – and how to set the stage of success. Covering wrong expectations, lack of engagement from older partners or GCS and fear of relocating jobs.

2 – Understanding and Demonstrating it for legal teams. Deep immersion in what he is (and is not) in the legal context, as well as concerns about credibility, ethics, hallucinations and explanation.

3 – Creating a change management book for legal technology. Speaking on how to design interested parties, lead the pilots, create champions, and structure a track in stages to capture hearts and minds.

4 – Approval of the direction and keeping it the moment. Displaying practical steps to build excitement, confidence and reliability over time.

And more….

Free free to attend but please RSVP here.

–