Should we use the AI in our case management system or a separate tool?

There is no single right answer, and the choice is less about the badge on the tool than about what it can see and what it does with it. AI built into a case management system has the advantage of already knowing your matters; a separate assistant is usually chosen because it does a specific job well. Judge both against the same questions: what documents it can read, whether every answer names its source, where your data is held, who inside the firm can see what, and how you pay.

What the built-in option offers

The case against adding another login is a fair one. A feature inside the system your fee earners already use has real advantages:

• No separate supplier to vet, contract with and monitor, if your existing agreement already covers it.
• No copying documents from one place to another, which is where confidentiality slips tend to happen.
• Access rights may already follow your matter permissions, so a fee earner sees only the files they are on.
• Training is shorter, because people are not learning a new product, only a new panel in a familiar one.

The limits are also real. A feature bundled into a wider platform may be narrow: good at summarising what is already in the system, less good at reading an official source end to end or showing its working on a calculation. It may be priced as an upgrade to a licence you must keep paying whether or not anyone uses it. And when a supplier builds everything, AI is one of many things competing for their attention.

Where a separate tool earns its place

A standalone assistant has to justify the extra login, the extra supplier assessment and the extra line in your register of systems. It usually does so on depth rather than breadth. The questions worth testing are practical:

• Can it work through legislation and official guidance at the source, rather than relying on a general summary of the law?
• Does it point at the page in the file, so a reviewer can check the answer against the document in seconds?
• Does it tell you when the papers do not answer the question, instead of producing something plausible?
• Does it show the steps in a date or figure calculation, each one carrying the rule it rests on?

If a tool does those things and the built-in feature does not, the second login buys you something. If it does not, you are paying twice for the same capability.

Questions to ask either supplier

Ask both the same list, in writing, and compare the answers side by side:

1. Where is our data held, and where is it processed?
2. Is our information kept apart from other firms, and is any of it used to train models for anyone else?
3. Inside our firm, who can see which matters? Does seniority alone grant a view, or does access follow the matter team?
4. Does every answer name its source, and can we open that source to check it?
5. What happens to a document page the tool cannot read: is it flagged, or quietly skipped?
6. What does the tool do on its own initiative? Does it send, file or serve anything, or does everything stop at a draft for a qualified person?
7. How do we pay: a subscription per user, a consumption model, a tier upgrade? What happens if we use it less than expected?
8. Can we export or delete our data, and what happens to it if the contract ends?

A built-in feature does not get an easier ride because you already trust the supplier for something else. Data protection obligations under UK GDPR, and your duties on confidentiality and competence, apply to the function, not to the invoice.

Running both without creating a mess

Plenty of firms end up with a case management assistant for routine work inside the system and a separate tool for research, analysis and calculation. That is a reasonable outcome, provided you decide deliberately rather than drifting into it. Write down which tool is for which job, so a fee earner does not have to guess. Make sure both appear on your list of approved systems, and that the file record of AI use covers whichever one was used.

Watch the cost shape. A per-seat subscription on top of a per-seat subscription is how AI spending gets out of hand in a firm of fifteen fee earners, particularly when half of them use it twice a month. If one tool charges only for what you use, it may sit comfortably alongside a bundled feature without adding fixed cost.

And set a review date. Both markets are moving. The feature that was thin at the start of the year may be solid by the end of it, and the standalone tool that justified itself in March may be doing work your case management system now does. Six months is long enough to learn something and short enough to change your mind cheaply.

Where Alesis fits

Alesis is an AI assistant for UK law firms, made by Alesis.ai Limited, used through the web browser with one conversation for a matter. It answers questions from the matter's own papers and names the page each answer came from, reads legislation, official guidance and Financial Ombudsman decisions at the source, and says so when the papers do not answer the question. The firm's information is held in the UK and processed only in the UK and the EU, kept apart from every other firm, and never used to train anything for anyone else. Alesis is funded by credit rather than a subscription, so it can sit alongside tools you already pay for without adding a recurring cost; it assists qualified professionals and does not replace them.