How to draft discovery responses faster without waiving objections (Texas and federal)

The response deadline and what missing it costs

Written discovery is one of the most predictable jobs in a litigation practice and one of the least forgiving. Under the Federal Rules of Civil Procedure, responses to interrogatories, requests for production and requests for admission are generally due 30 days after service, under Rules 33(b)(2), 34(b)(2)(A) and 36(a)(3). Federal requests for production delivered before the Rule 26(f) conference run from that conference instead. The Texas rules set the same 30-day default in Rules 196.2(a), 197.2(a) and 198.2(a), with 50 days for a defendant in a Family Code suit served before the answer is due. A stipulation or court order can change any of these dates, so the deadline for each set comes from the rule that applies and the orders in the case. This guide is general information, not legal advice, and the responsible lawyer must check the rules and orders that govern each matter.

The penalties are specific. A federal request for admission is admitted unless it is answered or objected to by the deadline. In Texas, Rule 198.2(c) says a late response means the request is considered admitted without the necessity of a court order. Any ground for objecting to a federal interrogatory that was left out of a timely objection is waived under Rule 33(b)(4), unless the court excuses the failure for good cause. Texas Rule 193.2(e) waives an objection that comes late or that is obscured by numerous unfounded objections, unless the court excuses the waiver for good cause shown.

The usual failure is timing. A set arrives, the associate is in depositions, the client is slow with answers, and the draft comes together in the last few days. An agent helps most at the front of the window, when the requests need sorting and the client needs a questionnaire. An early draft also leaves time to raise a dispute about scope with opposing counsel well before the deadline.

Where the hours go on a set

A typical set moves through several hands. An associate reads the requests and drafts objections. A paralegal turns the requests into a client questionnaire and chases the answers. The same paralegal Bates-numbers the production, and the associate then drafts the responses and, if anything is withheld, the privilege log. Each handoff is another chance for the set to sit in someone's queue.

On a contingency or flat-fee matter all of those hours are cost. On an hourly matter they are hours the client pays for and may question. Most of the time goes on reading requests and rewriting answers the firm has given before, which an agent does well. The judgment calls, such as which objections to make and what is privileged, stay with the attorney.

What to automate and what to keep with the attorney

The split below follows the rules. Anything the signature certifies, and anything that decides what the other side sees, stays with a lawyer. Products such as Briefpoint already sell discovery drafting, and Briefpoint says it drafts responses in your jurisdiction's format. A standard tool may be enough for some firms. A custom agent fits when you want your own objection bank and your own review steps in the loop.

Dates stay with your docketing system

The agent calculates a response date from the service date and method so the team sees the deadline on the first day. Your docketing clerk confirms that date, enters it in the rules-based calendar the firm already runs and owns any change from a scheduling order or a written extension. A missed discovery deadline can become a malpractice claim, so the agent works as a second check on the calendar and never as the calendar itself.

The client side of the clock needs the same early start. The agent drafts the questionnaire in the first days, along with a reminder schedule for the paralegal to send, so slow answers show up in week one instead of week four. When answers come back, it maps each one to the requests it covers and lists what is still missing.

Objections from your own bank

Boilerplate objections are a liability under both rule sets. Federal Rule 34(b)(2)(B) requires objections to requests for production to state the grounds with specificity, and Rule 34(b)(2)(C) requires an objection to say whether any responsive materials are being withheld on its basis. Texas Rule 193.2(e) waives objections buried under numerous unfounded ones, unless the court excuses the waiver for good cause. An agent that pastes the same twelve general objections into every response makes the problem worse.

The fix is an objection bank: the firm's own objections, each with the situations it fits and a short reason. The agent proposes objections request by request from that bank and shows why it chose each one. The reviewing attorney removes any that do not fit, and the bank changes only when the firm decides to change it. Writing the bank down also makes the firm agree on its positions, which helps every associate, with or without an agent.

Privilege logs and the signature

When a party withholds documents as privileged, Federal Rule 26(b)(5)(A) requires it to describe them in a way that lets the other parties assess the claim. Texas has its own procedure in Rule 193.3, built around a withholding statement. The agent can draft log entries from document metadata and content, but every privilege call is a lawyer's decision.

The signature carries the weight. Under Federal Rule 26(g)(1), an attorney who signs a discovery response certifies that it was made after a reasonable inquiry, and the client verifies the interrogatory answers. In Texas state court, an extension agreed with opposing counsel should be put in writing, because Rule 11 enforces agreements between attorneys only when they are in writing, signed and filed, or made in open court.

California's 2026 Practical Guidance names preparing discovery responses as an example of agentic AI. It says attorneys must not let such systems make substantive legal determinations or prepare and file pleadings without meaningful attorney supervision and review. Texas firms can use that as a working standard, since the agent drafts and a lawyer decides and signs.

Worked example: what the hours look like

This is an illustrative example with assumed numbers, not a client result. Take a litigation group of 12 lawyers that receives 5 sets of written discovery a week. Today each set takes about 6 associate hours and 3 paralegal hours, 540 minutes in all. Five sets at 540 minutes is 2,700 minutes, or 45 hours a week.

Assume 3 of the 5 are routine, such as a standard injury or contract case with familiar requests. With the agent, attorney and paralegal review takes about 240 minutes per routine set, so 720 minutes. The 2 complex sets, with heavy electronic discovery, stay at 540 minutes each, another 1,080. The total falls to 1,800 minutes, or 30 hours, which gives back 15 hours a week.

Timing matters as much as hours. An illustrative target is a full first draft by day 10 of a 30-day window, against an assumed day 25 today, which leaves room for the client and for any extension request. Your routine share depends on your practice, and a commercial litigation group may see fewer routine sets than an injury practice, so count a real month first.

Starting small with discovery

Start with the objection bank and twenty closed sets. Write down the objections your firm actually uses and when, then run an agent on copies of the closed sets and compare its drafts with what your team served. Count the objections the reviewer removed and the answers that needed rework. Use sets from matters that have closed, so the test never touches a live deadline, and run it inside the firm's systems, since closed files carry the same confidentiality duties as live ones. Ask the associate who served each set to review the agent's version, since they know which objections they dropped and why.

Our five-day AI Workflow Sprint does close to that, for $2,500 on redacted or synthetic samples, with the fee taken off the build. Full refund if we do not find 10 staff hours a week. Builds are from $7,500 per workflow, half to start and half when the acceptance criteria pass.

Common questions