Legal teams have invested heavily in tools to support contracting workflows. Yet in practice, many still operate with fragmented processes, inconsistent positions, and limited visibility across their agreements.
Across organisations, contracts are drafted, negotiated, and stored using a mix of templates, playbooks, emails, and institutional memory that rarely operates as a coherent system. Over time, this creates a familiar kind of friction: inconsistent positions, duplicated work, slow negotiations, and a persistent sense that legal is reacting rather than steering.
For years, legal technology has tried to address this by making the process faster.
But speed, as it turns out, is only part of the problem.
That is the starting point for SimpleDocs, an AI-native contract automation platform designed for in-house teams and law firms. But to understand where it fits, it helps to start with the person shaping much of its thinking: Electra Japonas, now Chief Product Officer at SimpleDocs, previously Chief Legal Officer at Law Insider, and co-founder of oneNDA.
Her career has not followed the typical legal tech trajectory of building tools around existing workflows. It has been more focused on questioning the workflows themselves.
Contracts Were Never Just a Drafting Problem
There is a tendency in legal tech to treat contracts as documents that need to be produced more efficiently. Electra’s view, shaped by years working in-house and across initiatives like oneNDA, is that the issue runs deeper.
Contracts do not just slow teams down. They introduce inconsistency, ambiguity, and decision fatigue across organisations. The same clause is negotiated repeatedly, positions drift over time, and institutional knowledge fragments.
From that perspective, automating drafting or redlining addresses symptoms instead of causes.
As she put it in the context of SimpleDocs’ development, the challenge is not simply adopting legal AI, it is grounding it in standards and internal precedent so that teams can actually trust and use the output.
That’s a subtle distinction, but one that changes how a product is built.
This week, The Legal Wire sat down with Electra for an in-depth conversation about how a career spent questioning legal workflows has shaped the way SimpleDocs is built.
TLW: You’ve worked across standardisation, contract data, and now product design. At what point did it become clear to you that the core issue with contracts was structure, not just speed?
Electra: “The shift happened when I stopped trying to make legal work faster and started asking why the same problems kept reappearing.
Early in legal ops, the working assumption was that contracts were slow because lawyers were slow. Add templates, add automation, add a CLM. But the bottleneck kept moving. Speed up drafting and the friction surfaces in negotiation. Speed up negotiation and it resurfaces in obligation management. The work compressed; the underlying chaos didn’t.
What became obvious, especially once I was working with AI on contract data, is that contracts aren’t slow because of effort. They’re slow because they carry no structure that anything downstream can read. Every contract is a private artefact, written for two parties, with no reliable way to find a clause, compare a position, or trust what the document actually says without re-reading it end to end.
Speed was the wrong frame. It was a productivity metric in a domain that had a data problem.
The product work I do now starts from that premise. Treat the contract as data first and a document second, and every other question becomes tractable: review, repository, intelligence. None of those work without structure. And no amount of model intelligence layered on top of unstructured legal text gives you a system you can actually trust.”
From Automation to Decision-Making
If drafting and redlining are no longer the hardest parts of contracting, what is?
Well, decision-making.
Legal work, particularly in contracts, is defined by trade-offs: risk versus speed, standardisation versus flexibility, precedent versus commercial reality. These are not problems that disappear with automation. These are the problems that need to be supported.
SimpleDocs is built around that idea. Its platform combines AI-powered drafting, redlining, and review with configurable playbooks and an AI-first contract repository, but the emphasis is less on generating documents and more on ensuring consistent, defensible decisions across a team.
Playbooks act as structured expressions of legal judgment. The repository turns past agreements into accessible knowledge. Benchmarking, which is powered by Law Insider’s vast dataset, provides context for what is typical or outlier.
Together, these elements aim to reduce one of the more persistent inefficiencies in legal work: not knowing whether the position you are taking is consistent with how your organisation has acted before. The design reflects an understanding that lawyers do not just need outputs, but a clear basis for the positions they take.
TLW: There’s a growing view that drafting and redlining are becoming commoditised. Where do you think the real complexity in contracting now sits and how should tools support that?
Electra: “The commoditisation framing is mostly correct, but it can be misleading. Drafting and redlining were never where the hard work actually sat. They were the visible part of the work, which is why people assumed that’s what lawyers did. AI is now exposing the difference between the surface task and the underlying judgment.
The real complexity sits in three places that have always been hard and are now harder to avoid.
First, position-taking. Knowing what to ask for, what to concede, what to fight, and why. This is organisational memory and commercial judgment. It does not live in the document; it lives in the people who have negotiated similar deals before, and in whatever record they kept of those decisions. AI doesn’t solve that. It makes the absence of a record more obvious.
Second, consistency. Most legal teams do not actually know what is in their contracts. Not in the redacted-PDFs-on-a-shared-drive sense, but in the structured, queryable, cross-deal sense. Once you can draft and redline at near-zero cost, the constraint becomes: are we taking consistent positions across the book? That’s a portfolio problem, not a document problem.
Third, obligations. The contract is signed and the work begins. Renewal dates, change-of-control triggers, audit rights, indemnity caps, data deletion duties. Almost every legal team operates blind on this and always has. Drafting was just the loudest pain.
So tools should stop competing on drafting speed. The frontier is the layer underneath: a structured, sourced, queryable account of what your contracts say and why they say it. Drafting and redlining sit on top of that layer. Without it, you are speeding up the easiest part of the job and pretending it’s the hardest.”
Data, Precedent, and the Law Insider Effect
The merger between SimpleDocs and Law Insider in 2025 was not just a strategic expansion. It was, in many ways, a structural shift.
Law Insider brought with it a dataset of more than 5 million contracts and 20 million clauses across 50+ languages, providing something most legal AI tools don’t have access to: real-world precedent, at scale.
This type of dataset reflects how contracts are used in practice, and this is crucial because contracts are inherently comparative. Lawyers rarely (if ever) ask what the law says in abstract. They ask, what is acceptable here? what has been agreed before? what is market?
By integrating this dataset into its platform, through features like benchmarking and the Law Insider Index Score, SimpleDocs is able to ground its outputs in something closer to practice than theory.
It also aligns with Electra’s long-standing focus on standardisation. OneNDA demonstrated that shared standards can massively reduce friction, while Law Insider showed that real-world data can inform those standards. SimpleDocs brings those ideas into a system that can be applied day-to-day.
TLW: Law Insider introduced large-scale access to real-world contract data. How does that kind of dataset change what is possible when designing contract automation tools?
Electra: “Data on its own changes less than people assume. The legal industry has had access to large volumes of contract text for decades. SEC filings alone are an enormous corpus. What was missing was not the data. It was the structure that made the data useful, and the tools that could query it at the level of clause, position, and pattern rather than document.
What Law Insider did, and what becomes more powerful inside a contract automation system, is collapse the distance between a question a lawyer wants to ask and an empirical answer. Three things shift when you actually have that.
First, position-taking becomes evidence-based instead of folklore. Most negotiation positions are inherited. “We always ask for a 12-month survival period.” Why? Because the partner who trained you said so. With real-world data at scale, you can ask what survival periods actually look like in this industry, this deal size, this counterparty class. The answer often disagrees with the folklore. That is uncomfortable for the profession and useful for everyone else.
Second, provenance becomes possible. AI that generates contract language without grounding produces plausible text that may or may not reflect any position anyone has actually taken. AI grounded in a corpus of real contracts can show you where the language came from, how common it is, and how it has been negotiated historically. That changes what trust in an AI output looks like. It moves from “the model said so” to “here is the source, here is the precedent.”
Third, you can train models that actually understand contracts. General-purpose LLMs are trained on the internet. Contract language is a distinct dialect with its own conventions, drafting patterns, and unwritten rules. A model trained or grounded on real contract data has a fundamentally different competence on legal text than one fine-tuned on generic instructions.
The caveat is the one I keep coming back to. Data alone does not give you a system. You also need structure: a way to ask the question, a way to map a clause in the corpus to a clause in the deal in front of you, a way to surface the answer at the moment of decision. That is the design problem we are working on. Law Insider gives you the empirical layer. The contract automation tool has to make it usable inside the workflow, otherwise what you have is a research database, not a system.”
Contracts as Systems, Not Documents
An important aspect of SimpleDocs is how it treats the contract lifecycle as a connected system rather than a series of steps.
Drafting and redlining happen within Microsoft Word, integrated into existing workflows. Playbooks ensure consistency across teams and the repository turns contracts into searchable, structured knowledge. Workflow automation and the “legal front door” manage intake, routing, and approvals.
Individually, none of these ideas are entirely new. But together, they form something cohesive.
The aim is not only to move faster through each stage, but to reduce the friction between them.
That is an important take-away for many, especially large organisations, where contracts are less about individual documents and more about how consistently and predictably the organisation operates at scale.
TLW: You’ve spoken about moving from documents to data. What changes in practice when contracts are treated as part of a connected system rather than standalone agreements?
Electra: “The shorthand answer is that things which used to be impossible become routine, and things which used to be routine become invisible.
The longer answer needs an example, because the practical change is not obvious until you see it.
Take a normal scenario. A large customer asks for a contract amendment. Before contracts are connected data, the legal team’s first question is: what does the existing contract say? Someone opens a PDF. Reads it. Tries to remember whether similar amendments have come up with other customers. Maybe emails two senior colleagues. Maybe searches a shared drive. Whatever they decide gets buried back into a new PDF and the cycle resumes.
In a connected system, the question changes. The lawyer can now ask: what does this contract say, what have we agreed in similar amendments across the book, where have we held and where have we moved, what is our aggregate exposure if we accept this one. None of that requires reading. It requires querying.
That single shift cascades into several practical changes.
Routine work disappears as work. Most contract questions inside a business are not legal in any deep sense. They are factual. What does the contract say, when does it renew, what notice period do we need to give. In a connected system those are answered by the system, not by a lawyer. Sales can answer its own questions. Finance can answer its own questions. Legal stops being the bottleneck for things that were never really legal work.
Risk gets re-mapped. Most legal teams operate on an intuition of where their exposure lies. Once you can query the contract estate as a portfolio, the intuition turns out to be wrong roughly half the time. The contracts that worry the GC are not always the ones carrying the real exposure. The exposure tends to live in the long tail of overlooked, mid-size agreements that nobody thinks about until something goes wrong.
Negotiation becomes self-correcting. Position-taking that previously relied on whichever senior lawyer happened to remember similar deals now relies on the data. Outliers stand out. Drift stands out. A junior lawyer can see what positions the team has actually held, not what they think they have held.
Obligations stop being a discovery exercise. Most legal teams find out about an obligation when something has already gone wrong. In a connected system, obligations surface at the moment they matter: before the renewal, when a regulator changes a rule, when a counterparty triggers a clause.
The most underrated change is cultural. When legal is no longer the gatekeeper for facts about contracts, the role changes. The team stops being asked basic questions and starts being asked harder ones. That is uncomfortable for some lawyers and liberating for others. In my experience, it is also the real reason legal teams resist this shift. Most professional resistance to data-driven legal work is, at root, resistance to a change in what legal is for.”
Augmenting Legal Judgment
A recurring theme in discussions around AI in law is the potential replacement of lawyers.
SimpleDocs takes a more grounded approach. Its design suggests a focus on augmenting legal judgment, instead of bypassing it. Playbooks encode how a team thinks, AI applies those rules consistently, and data provides context, while lawyers remain responsible for interpretation, negotiation, and final decisions.
To us, this seems like a realistic division.
Legal work has always combined structured reasoning with contextual judgment, and technology can support the former more easily than the latter. The value lies in reducing the cognitive load around routine decisions, allowing lawyers to focus on the areas where judgment is genuinely required.
TLW: Where do you think legal judgment should remain irreducibly human, even as systems like SimpleDocs become more capable?
Electra: “The honest answer is that the boundary keeps moving, and people who claim to know exactly where it sits are usually defending whatever they currently do.
That said, three categories feel durable.
First, defining what good looks like for the deal. AI can negotiate against a playbook. It cannot tell you whether the playbook reflects the right commercial position for this counterparty, this market, this moment. That judgment requires context the system doesn’t have: the relationship, the politics, the strategic frame around why you are signing this contract at all.
Second, accepting responsibility. Even if a system produces a defensible recommendation, someone has to own it. Not legally, in the malpractice sense, but professionally. A lawyer signs off. A board signs off. AI cannot. As tools get better, the volume of decisions a single lawyer is accountable for goes up, not down..
Third, setting the standards the system runs on. Every AI contract tool is downstream of decisions about what positions are acceptable, what exceptions are tolerated, what risk is worth fighting. Those decisions are made by humans, often badly, and they are the highest-leverage judgment in the entire system. A team with a thoughtful playbook and a mediocre AI will outperform a team with a brilliant AI and no point of view.
Where judgment is not irreducibly human is in the bits people have romanticised: clause-level review, position comparison, drafting against a known rule. That work was never the judgment. It was the labour around the judgment. The mistake is thinking the labour and the judgment were the same job.”
Where SimpleDocs Fits
SimpleDocs does not attempt to redefine contracts entirely. It operates in a more pragmatic space: improving how they are created, negotiated, and managed within organisations that already depend on them.
Its differentiation lies less in any single feature and more in how those features are connected and in the thinking behind them.
That thinking is shaped by a career that has moved through standardisation, data, and now systems. The product reflects that trajectory.
For legal teams, particularly those operating at scale, the question, instead of simply being whether contracts can be automated, is shifting to whether they can be made consistent, understandable, and aligned with how the business actually works.
SimpleDocs offers a well-developed response to that question.
TLW: You’ve worked across in-house roles, standardisation initiatives like oneNDA, and now product design. Looking back, what do you think the legal industry has consistently misunderstood about contracts and why has that misunderstanding persisted?
Electra: “The consistent misunderstanding is that a contract is a document. Almost every assumption the legal industry operates on flows from this. That contracts are bespoke. That they are best produced by skilled drafters working from precedent. That quality is measured at the level of the individual artefact. That value is delivered when the document is signed.
None of that is wrong, exactly. It is just the wrong unit of analysis.
A contract is more usefully thought of as a structured record of a commercial relationship that happens to be expressed as prose. The document is the rendering, not the thing. What matters for the business is the data the contract carries: parties, obligations, dates, financial exposure, termination rights, risk allocation. The fact that this data is encoded in long-form language is an inheritance from a pre-digital era, not a feature.
Once you accept that framing, most of the standard practice looks slightly absurd. You spend hundreds of hours drafting a unique document for what is, ninety percent of the time, the same commercial relationship as the last twenty deals. You store the result as a PDF you cannot query. You then commission another set of lawyers to read the PDF when something goes wrong. The waste compounds.
The reason it has persisted is partly economic and partly cultural. The economic part is that the billable hour rewards artefact production, not system design. There is no margin in standardising what you can charge to draft from scratch. The cultural part is that the legal profession trains people on individual matters and judges them on craft. There is no career path for the lawyer who reduces ten thousand bespoke contracts to fifty good templates.
Initiatives like oneNDA worked precisely because they sidestepped both incentives. They were built outside firm economics, by people who had stopped pretending the bespoke version added value. That is the model. The shift from documents to data needs to happen the same way: from outside the structures that benefit from the current confusion.”
TLW: Given your experience across both legal practice and product, how do you think the role of in-house lawyers will evolve as contract systems become more structured and data-driven? What will good legal judgment look like in that environment?
Electra: “The in-house role is going to bifurcate, and the legal industry has not yet adjusted to that.
The old version of the role was, in practice, a high-volume execution job dressed up as advisory work. You reviewed contracts, you negotiated NDAs, you triaged commercial requests. You called yourself a business partner, but most of your time was spent on document throughput. As contract systems get more structured, that throughput layer disappears. Not in five years. It is already disappearing.
What replaces it is a smaller number of higher-leverage roles built around three things.
First, designing the system. Someone has to decide what the playbooks are, what positions the company will take, what exceptions get escalated, and how the contract estate is structured for query. This is product work, not legal work in the traditional sense, but it requires deep legal judgment. It is also the most important thing the in-house function can do.
Second, handling the exceptions the system surfaces. Once routine work is automated, what remains is the genuinely hard stuff: novel deals, contested positions, regulatory ambiguity, deals where commercial context overrides the standard answer. This is where lawyer-as-judgment lives. It is a smaller volume of work, but it is the work that justifies the role.
Third, owning the portfolio view. In-house teams have historically operated deal by deal because that is how the work arrived. Structured contract data makes it possible, for the first time, to ask portfolio questions. Where is our exposure concentrated? Are we taking consistent positions? What changed across the book this quarter? Good legal judgment will increasingly mean answering those questions, not just the question in front of you.
The lawyers who do well in this environment will be the ones who treat the contract function as infrastructure rather than service delivery. That is a different professional identity. It involves losing some of what the profession has historically valued, particularly the craft of bespoke drafting, and gaining things the profession has historically undervalued, particularly systems thinking, data fluency, and a willingness to standardise.
Most in-house teams are not set up for this transition, and most law schools are not preparing people for it. That gap is the real story of the next five years.”
