Every legal team has heard it. "Why is this deal not signed yet?" It lands in a Slack DM, in a hallway, in a Monday-morning email with three exclamation points. And most of the time, the answer isn't as simple as "legal is slow."
We got a room of people who spend their entire professional lives inside this problem, Akshay Verma COO at SpotDraft, Pralika Jain, Head of Legal at Hex, Jonathan Perkel, CLO at Esusu, and Alicja Kwiatkowski, Senior Counsel EMEA and Group Data Protection Officer at Mirantis, and asked them to help us investigate during our recent webinar; "The Case of the Slow Contract: An AI-Era Investigation," in partnership with In-House Connect . Here's what the case file turned up.
The instinct, when a deal stalls, is to assume legal is the bottleneck. The panel's collective response: wrong lead.
Slow contracts are almost always a people-and-process problem, not a legal-judgment problem. A sales rep, a marketing procurement lead, and a tech procurement lead are three completely different customers with three completely different needs — and treating them with a one-size-fits-all process is where things start to break down. Build the process around the actual audience first, then let the people slot into it.
If you've been in a procurement negotiation in the last two years, you've felt this one. AI clauses went from a non-issue to the single biggest source of friction almost overnight, a 2025/2026 problem that actually started brewing in 2024.
The questions are always some version of: What do you do with our prompts? How is our data processed? Who are your sub-processors? These aren't unreasonable questions, procurement teams are right to be curious. The fix isn't to dodge them, it's to build a repeatable, self-serve answer pattern so the same five questions don't reinvent the wheel every single cycle.
Here's the uncomfortable truth the panel surfaced: most legal tech is built to serve legal. But contracting isn't a legal-only process, it touches revenue, risk, privacy, finance, and legal all at once. So who owns it?
"It depends" isn't a cop-out here, it's genuinely the answer. It depends on your org structure, your maturity, your risk appetite. One panelist described a deal that had to clear IT security, then wait on finance to confirm budget, and only then land on legal's desk two weeks later, at which point everyone was suddenly furious that "legal" hadn't signed it yet. The real fix isn't more approval gates. It's one platform, with the right permissions and AI-driven routing, so IT security, finance, and legal aren't three separate silos a contract has to survive.
Every contract has two lives: the time someone is actively working it, redlining, negotiating, approving, and the time it just... sits. The panel's read: that split isn't fixed, it's a function of what matters to the person holding it up.
A $24,000 deal with a recognizable logo might sit untouched because, in someone's head, the dollar figure doesn't justify urgency, even though the strategic value is high. The move here isn't to complain about the delay. It's to build internal clarity on what actually counts as urgent, and to learn the specific patterns of the people you're waiting on. One genuinely useful hack from the conversation: give a deadline and a default. "I'm sending this to the customer by Thursday unless I hear otherwise." Calculated forgiveness beats endless permission-seeking.
Every lawyer has been the associate who marked up an entire contract just to prove they were paying attention. The panel was refreshingly honest about this: over-redlining doesn't generate value, it generates cycles. The skill that actually matters is knowing which three issues in a contract are the ones worth fighting for — cap on liability, indemnity, the usual suspects and using comments instead of blanket edits so the back-and-forth stays legible instead of becoming an archaeology dig.
This is where the conversation got most interesting. AI is genuinely reshaping contract review, flagging gaps, summarizing redlines, cutting a 20-minute security-document review that used to eat an afternoon. But the panel was clear-eyed about the catch.
There's a hidden tax that comes with the speed: expectations. The moment a team adopts AI, the assumption becomes "this should be instant now" and that's not how it works yet. The phrase that stuck from the session: "meat in the loop." Not "human in the loop" as a checkbox, but an actual person reviewing whether the AI's flags make sense, which, done properly, can eat up nearly as much time as doing the review manually in the first place.
And there's a deeper irony sitting underneath all of it: freeing up time via AI doesn't necessarily mean lawyers get more strategic headspace. It might just mean the same lawyer now processes 60 contracts a month instead of 20. Capacity gets absorbed, not necessarily redirected toward higher-value work, something worth naming out loud before you build your team's AI roadmap around "we'll finally have time for strategy."
A pattern the panel flagged without much prompting: business stakeholders are increasingly showing up with "I asked ChatGPT and it said X, can you validate?" One panelist shuts this down fast, general-purpose AI answers without full business context are often generic, and there's a real privilege question buried in whether those AI conversations are even protected. The guidance: stick to giving legal the facts, not the AI-generated pre-analysis, and build clear internal lines for what's safe to send externally and what isn't.
Nobody on the panel pretended contracting is about to become instant. But the throughline across every "suspect" was the same: the fix is rarely a technology purchase alone. It's process clarity, ownership clarity, and a willingness to get comfortable with a level of risk that used to feel uncomfortable. As one panelist put it about an over-negotiated influencer agreement for a single photo post: sometimes the real question is just - what's the worst that can happen?
That mindset shift, more than any single tool, is what actually closes the case.
This recap is based on SpotDraft's webinar "The Case of the Slow Contract: An AI-Era Investigation," featuring Akshay Verma, Pralika Jain, Jonathan Perkel, and Alicja Kwiatkowski.
