Practitioners at IP Counsel Cafe this year broadly agreed that the nearer-term shift is less about work allocation and more about how in-house teams and outside counsel work together, with clearer standards, higher quality, and more consistency on both sides. The common thread is how the work gets specified. Getting higher quality and efficiency from AI means stating clearly what good work looks like in a form a tool can apply and review. That is codification, and it matters far more now than it used to.
None of this assumes AI is close to replacing professional judgment. Reasonable practitioners may disagree about how capable it will become, and many doubt it will match seasoned judgment for years. The case for writing practice down does not rest on any forecast about AI. It rests on something already visible, namely that today's systems can follow clear instructions that reflect how strong work gets done.
Patent work has long run on expertise that is easier to exercise than to write down. An experienced practitioner knows how to structure a claim set, where a specification needs depth because of how prosecution tends to unfold, and when an office action calls for argument rather than amendment. The same holds for in-house and outside counsel alike. Many teams keep procedural checklists for filing and docketing, but the more substantive parts of the craft tend to stay with individual practitioners and in the work they have produced, which is largely where junior practitioners learn them.
There are understandable reasons for this tendency. Until recently, the return on setting out how one approaches substantive issues was modest, mainly some personal consistency and better training material, and that effort competes directly with billable work or other higher-priority tasks. The subject also resists easy organization, since substantive judgment depends on the technical field, the jurisdiction, business strategy, and a body of law that keeps changing, often at once. Seen that way, not investing heavily in writing it all down may have been a reasonable choice.
Take the drafting of a set of claims for a complex genus. No written rule tells you how many dependent claims to spin out, yet the total has to stay manageable rather than mechanically calling out every alternative in its own dependent claim. Over time, a practitioner develops a feel for it, promoting the alternatives that matter, such as those covering the exemplified or commercially significant embodiments, and folding the rest together. That feel is genuine expertise, and it usually stays unwritten. The moment you try to tell an AI how to draft those claims, you have to put into words the unspoken heuristics you had been applying.
AI changed the payoff for codifying the practice. A codified standard used to do its work only when a colleague read it. Now the same standard can guide AI-assisted drafting on new matters and inform the review of the drafts that come back, across a whole portfolio. In effect, one practitioner's hard-won expertise, once codified, can raise the drafting and review of an entire team rather than staying locked in one person's matters. Written or codified practice has gone from a training aid to something that shapes output directly at scale.
What AI has not changed is who is in control. Codified standards guide rather than dictate. They can hold open-ended considerations as readily as fixed rules, and they stay easy to revise. So the practitioner, not the tool, keeps the final call. The same holds for accountability. A named professional still signs off on the filing and stands behind it, whatever tools produced the draft, so the judgment that shapes the result stays human even as the drafting is shared with a machine.
This is where the tool itself matters. Solve Intelligence is designed to keep the practitioner in command, with clear visibility into what the AI produces at each step and the underlying reasoning, and the control to change any of the output. Used this way, codification extends an expert's reach across a portfolio and frees their judgment for the highest-level decisions, which stay firmly with them.
Anyone who uses AI at all builds on some foundation, and the question is how much of it a team supplies itself. The purpose-built platform can provide the out-of-the-box codification of best practice, the patent data, and tooling around it, and the work of keeping all of it current as models, the law, and practice change. That work is substantial and never finished, yet none of it sets one team apart, so building it yourself spends scarce time you could put into what does. The more of the foundation a team starts with, the more of its remaining time goes to its own codification and judgment.
A frontier model on its own gives you a thin foundation. It is a genuine starting point, but a general one. On a given matter, it may not reliably reflect current practice in a specific office or jurisdiction. It has no reliable access to patent data such as prior art and litigation history. It has no native handling for figures, chemical structures, or sequence listings, and no one keeps it current for patent-specific correctness. A team that starts there is choosing to supply the patent layer itself.
A platform like Solve Intelligence supplies far more of that foundation. It sits on top of the same frontier models and adds the patent layer, built and maintained by patent professionals, so a great deal of best practice is already encoded out of the box and a team starts from a higher baseline. Because that foundation is shared across more than 700+ IP teams, it is kept current far more efficiently than any single team could manage alone.
That foundation shows up in concrete features. Auto Template for Solve’s Drafting, Prosecution, and Charts products turns a team's existing filings and internal standards into custom AI templates, so a team's own codification starts from real work rather than a blank page. That higher-level codification stays with the team. Solve Intelligence gives it a private, maintained place to encode and apply its custom AI templates and instructions, and the team decides how much of its judgment to write down and how much to keep in its people's hands. Where a team wants help capturing its own practice as standards, Solve's practitioners can assist, and the substance and its confidentiality remain the team's. The LexisNexis integration puts global patent litigation data inside every workflow within Solve’s platform, so codification and everyday decisions rest on current data.
A good place to start is what the purpose-built platform already provides. Because it encodes a great deal of common best practice, a team can run on that foundation and add none of its own codification, applying its people's judgment to each case, and for many teams that is the right amount. The open question is how much of its own to add on top. A team's own codification can capture what is distinctive about its practice and make it consistent across its people. But whatever a team adds is its own to build and keep current. The choice, then, is how much to codify versus how much to leave to the platform and handle through everyday judgment. That balance is a moving target, and each team reassesses it as its practice, the law, and the technology change.
Some of a team's own codification is worth it for almost everyone, such as a custom application template or a house style captured once and reused. The larger decision is how far beyond that to go, toward the higher-level standards that carry more of a team's judgment and need more upkeep. An outside-counsel group, for instance, might codify how it structures claims and specifications to withstand the validity challenges it keeps seeing in its field. An in-house team might codify what a specification has to cover for its own products and standards, so every outside firm starts from the same baseline. Whatever a team builds this way is its own, and often confidential.
The same question runs through the relationship between in-house teams and outside counsel. An in-house team might codify enough to run a first-pass review of incoming drafts, or to hand outside counsel a structured starting point to build on. A law firm might codify its practice-group standards to make its edge portable, or keep that edge in its partners' judgment. Neither side has a settled playbook yet, but the direction is becoming clearer. With the shared foundation handled, the advantage goes to whichever side puts the time it saves to better use. That could mean sharpening its strategic judgment, codifying its own expertise, or both, then turning the result into consistent quality across everything it drafts and reviews.
Start with the decision that is already clear. Get onto a strong, well-maintained foundation rather than rebuilding one yourself. Then treat how much of your own to codify as an ongoing experiment, not a formula. For in-house teams, the near-term step may be to add the codification that clearly pays off, to be deliberate about how much further to go, and to decide what to build versus what to leave to a partner. For outside counsel, codifying the expertise of a practice turns it into something the law firm can demonstrate and keep as its own. Keeping those standards sound as the law develops is exactly the kind of judgment that stays valuable. For both, the strategic move is to put scarce effort into the codification and judgment that differentiate the work, and to let a partner carry the foundation.
Solve Intelligence is built for either side of the table. A law firm can hold its own codification on the platform as its asset, and an in-house team can maintain and refine its standards there. But codification is not the only thing that stays with the team. However the work is divided, and however much is handed to the tool, the responsibility for the result does not move. That is why the judgment behind the standards remains the practitioner's to own.
Solve Intelligence gives your team a practitioner-built foundation out of the box and a private place to codify the expertise that sets you apart, kept as your own. To see how in-house teams and outside counsel are putting it to work, book a demo or talk to our team.
Codification means turning the way a team works into templates, instructions, and review criteria that an AI tool can apply directly, rather than knowledge that lives only in a practitioner's head or in past work product. It can range from a house drafting template to higher-level standards for how claims are structured in a given field.
Not necessarily. A purpose-built platform already encodes a great deal of best practice out of the box, so a team can run on that foundation and apply its people's judgment to each case. Adding your own codification on top is an option that may pay off for your team, not a prerequisite.
Because the groundwork under any codified standard is substantial and never finished. It covers current patent data, tooling, and keeping everything correct as models and office practice change. Every team needs that work, but none of it necessarily sets one team apart, so building it yourself spends scarce effort on what differentiates no one. A platform does this heavy, shared work once and keeps it current for many teams, far more thoroughly and efficiently than any single team could alone, so your scarce effort goes into the codification and judgment that actually differentiate you, which stay yours.
The nearer-term shift is less about work allocation than about how the two sides collaborate, with clearer standards, higher quality, and more consistency on both sides. In-house teams can hand outside counsel a structured starting point or run a first-pass review, while outside counsel can make their expertise more portable and demonstrable.
No. A named professional still signs off on a filing and stands behind it, whatever tools produced the draft. Codification changes what a team can hand to the tool, never the responsibility for the result.
