Your Discovery Data Doesn’t Live Alone Any More

A legal matter is a business problem moving across systems: scope changes, collections expand, review decisions accumulate, invoices land, deadlines shift, and strategic calls get made in emails, meetings, and project plans. The challenge is managing that business record as one connected picture, even when the information lives in many places.

Moving fact development earlier was the argument in my first blog post on Agentic AI. The next opportunity is broader: using agents to reason across the business of the matter itself. Not just what the documents say, but what the work is costing, what scope is changing, what decisions have already been made, and where burden or risk is beginning to concentrate.

That is the problem agents can help solve. The real corpus is the matter: documents, review calls, production history, collection inventories, budgets, invoices, staffing plans, deadlines, client context, and prior matter knowledge. The value is not pulling everything into one platform. The value is giving counsel and legal operations a controlled way to ask matter-level questions across all of it.

A repeatable playbook gives that work a process. Now consider what the process needs to execute. A rule that says to revisit collection scope when a new business relationship appears needs both the collection history and the client’s organizational context. A rule that says to assess what another round of investigation is buying needs findings and spend. The instructions can be sound while the information needed to follow them sits somewhere the agent cannot reach.

People fill that gap today. The review team knows the coding. Legal operations has the billing records. Someone at the client knows that the project changed names after an acquisition. The case team connects those pieces through experience and a fair amount of coordination. Asking the same question again means repeating much of that work.

This is where Model Context Protocol, or MCP, becomes relevant. It gives AI applications a common interface for tools and information exposed by connected systems. Where approved connections exist, an agent could consult the discovery platform, request information from another system, and use analytical tools to work through the question. The database can retain its review and production responsibilities while contributing to decisions outside that workflow.

Consider a team deciding whether another collection is worth the proportionality argument. An agent could compare the proposed source against prior collection history, review yield, issue coverage, privilege burden, staffing assumptions, budget burn, and upcoming deadlines. The useful answer is not simply “collect more” or “do not collect.” It is a reasoned view of whether the additional work is likely to change the matter posture, reduce risk, or simply add cost.

That requires more than search. Review metrics may show what surfaced. Billing may show what the work cost. Project plans may show why the team staffed the matter a certain way. Client records may explain why a source matters at all. The agent’s job is to reason across those records while keeping the source, confidence, and limits of each conclusion visible.

This is what gives the connection value over the life of the matter. A new collection affects scope. Review activity affects staffing and cost. Issue coverage affects negotiation strategy. Budget burn affects the appetite for the next step. The playbook defines when to revisit the question and when a person must decide; the connected sources provide the business record needed to make that decision intelligently.

Which brings us back to choosing the first workflow. We’ve already established that we should not always start automation with production decisions. That discipline matters more when an agent could reach across systems. There is no reason to request broad access just to discover that two exports cannot be reconciled reliably or that nobody can define what the answer should mean.

For the example above, begin with an approved collection inventory and the relevant organizational records. Test whether the process can identify a coverage question, explain its basis, and recognize when it lacks enough information. Leave the collection decision with counsel. If the process holds up, a narrow live connection can address the burden of refreshing it. Add billing only when there is a defined question it can help answer.

We do not need every system connected to learn whether one connection is worth maintaining. We need a question that matters often enough to justify the work and a process the case team can inspect.

Agents become valuable when they can reason across the whole business of a matter, wherever that record lives. The next collection should be evaluated against what prior work revealed. The next spending decision should reflect what the last round actually bought. We are already paying to create this intelligence. We should not have to reconstruct it every time the matter moves.