SystemsPERSUIT2026

Designing an AI agent lawyers would actually trust

Persi advises Fortune 500 legal teams on whether to keep a matter in house or send it to outside counsel. I led design from experimentation to beta — principles, content design, research, and the system components.

My role: Design lead — principles, content design, user research, design system
Focus: Product design, Design system, Strategy, AI
Outcome: 22 beta customers; a deciding factor in contract renewals

The Persi interface shown on a laptop against a dark office background, with the conversation on the left and an expanded Business Rules artifact panel on the right

Context

Persi is PERSUIT’s legal operations agent. It helps Fortune 500 legal teams decide when to keep a matter in house versus hire outside counsel — weighing subject matter, size, geography, and cost.

I joined as design lead during experimentation, while the model and data architecture were still being worked out. Persi had to sit inside a platform users already knew: enough cohesion to feel like PERSUIT, enough distinction to signal a genuinely new kind of experience.

The key decision

Where the agent hands off to a human. An agent that never hands off frustrates people exactly when stakes are highest. One that hands off too readily never earns trust. Setting that line shaped everything downstream.

What I did

Grounded the team in evidence before setting principles. I ran competitive analysis across adjacent software and leading AI chat products, then shared it out. Without shared reference points, principle-setting becomes taste, and taste doesn’t scale.

Competitive discovery research board comparing six AI products, with screenshots grouped by pattern and annotated in pink, alongside a highlighted key insight panel — Six products pulled apart by interaction pattern.

Ran three workshops with 11–13 participants — PMs, engineers, architects, designers — covering memory, human intervention, and navigation between surface areas.

The resulting principles.

Introduced a palette and new components, evaluated with design system engineering leads for scalability and maintenance cost. Agent work surfaced needs the existing system didn’t have, so each addition had to earn its keep.

Design system proposal board showing colour swatches alongside proposed Persi components — message states, artifact panels, warnings, and result cards — laid out for team review — What shipped was a compromise between this and the legacy system.

Specification for the prompt box component showing its states across empty, focused, filled, disabled, and error, with annotations for each — The one component every use case passes through, so it had to survive all of them.

Worked across every squad’s use cases — matter intake, governance approvals, e-billing status. Patterns that hold for an invoice check fall apart on a governance approval, and you only see that with all of them in view.

Persi user flows organised by squad, with a horizontal lane per squad and each step captured on a yellow card across panel management, data, firm pricing, e-billing, and intake

Grid of condensed chat states across different squads and phases, showing how a Persi conversation appears at each stage of several different use cases — The same conversation model, stress-tested across use cases.

Rules and governance screen where a user edits business rules in plain language on the left and Persi generates the structured matter routing, RFP creation, and in-house handling rules on the right — The user writes the rule in their own words; Persi returns the structured version for approval.

De-risking through content design

When an agent tells a general counsel that benchmark fees are €380K–€520K, the words around that number decide whether it gets acted on, ignored, or escalated. Content design is the trust surface.

Persi confirming a set of business rule updates, with the conversation on the left and an artifact panel on the right listing the new RFP requirements, in-house handling rules, and billing and engagement terms, with a collapsed warning at the top — Changes summarised in an artifact panel rather than buried in conversation, so what the agent committed to is reviewable in one place.

The same artifact panel with its warning expanded, listing the specific rules the user must confirm before the changes take effect, each stated in full — The warning states every rule in full. Making the user open it is friction on purpose — this is where an unnoticed mistake becomes a governance problem.

Outcome

22 customers on the Persi beta program
A deciding factor in renewing their contracts with PERSUIT

Research sessions with a subset showed they wanted benchmarking data faster — and a partner in the in-house versus outside-counsel decision, not a system that made it for them.

That validated the intervention line. The value wasn’t automating judgment. It was removing the work getting in the way of it.