An auto-renewal is a decision made by forgetting

Why renewals are lost quietly

Nothing dramatic happens. A date passes, the term extends on the vendor's terms, and the organization finds out when the invoice arrives.

• A spreadsheet and good intentions: Notice periods are read off a PDF once and typed into a tracker
• A spreadsheet and good intentions: The tracker stops being maintained when its owner changes role
• A spreadsheet and good intentions: A renewal surfaces after the window, when the only option is to accept
• A spreadsheet and good intentions: Nobody knows whether the agreement auto-renews or simply ends
• A spreadsheet and good intentions: Price increases are discovered at invoice rather than at negotiation
• A spreadsheet and good intentions: Amendments change the notice terms and the tracker never learns
• Obligations extracted from the paper: Notice windows and renewal triggers come from the executed document
• Obligations extracted from the paper: The record does not depend on anyone maintaining it
• Obligations extracted from the paper: Each window surfaces with enough runway to actually negotiate
• Obligations extracted from the paper: Auto-renewal versus fixed expiry is a known property of the agreement
• Obligations extracted from the paper: The renewal conversation starts before the price is a fait accompli
• Obligations extracted from the paper: An amendment updates the operative terms, including the dates

What happens when a window opens

Renewals are bounded and repetitive, which makes them a natural thing to delegate. You set the policy once and decide only on the exceptions.

• The window surfaces itself — Because the notice period was extracted from the executed agreement, the renewal appears with runway rather than as an invoice surprise.
• The agreement is summarized — Current terms, pricing, what was negotiated last time, which clauses deviate from your standard and what the vendor conceded before.
• A recommendation, with reasons — Renew as-is, renegotiate, or let it lapse — argued from the agreement's own terms and your policy rather than asserted.
• The conversation runs inside policy — Where you choose to renegotiate, the assistant opens it, proposes your position, handles the vendor's response and escalates anything outside the mandate.
• The outcome becomes a version — A renewal or amendment attaches to the clauses it changes, so the operative terms and the next notice window update with it.

Renegotiation is a timing problem first

The best renewal position is held by whoever opened the conversation early. Everything the assistant does here is in service of starting while you still have an alternative.

• Early, not urgent — Windows surface on the notice period in the document, so the conversation starts when lapsing is still a credible option.
• A policy per vendor tier — Acceptable uplift, term length and payment terms by spend tier, so the assistant knows the difference between a routine renewal and one that needs attention.
• Concentration and exposure visible — Which clauses deviate from your standard, where coverage or liability terms sit, and what the agreement commits you to beyond price.
• Escalation with the history attached — When an approver is needed, they get the current terms, the proposed terms and the last negotiation's outcome in one view.

Set once, then stop thinking about it

The configuration is small because the extracted obligations do the work. You are describing tolerance, not building a tracker.

• Policy: Acceptable uplift by spend tier
• Policy: Preferred term length and payment terms
• Policy: Clauses that always require an approver
• Policy: Turn and spend caps per renewal
• Timing: How much runway a window should surface with
• Timing: Different lead times for different spend tiers
• Timing: Evergreen agreements treated as recurring decisions
• Timing: Amendments update the next window automatically
• Ownership: A business owner per vendor relationship
• Ownership: Approvers by spend threshold
• Ownership: Escalations arrive with the agreement's history
• Ownership: Organization-wide kill switch for all agents

What teams ask about renewals

• Does this work for agreements we signed before using Paqt? Yes — that is the usual starting point. An executed agreement brought into Paqt has its terms and dates extracted the same way, which is often the fastest way to find the renewal windows nobody is currently tracking.
• Can the assistant renew something without asking us? Only within the policy you set, and only for the tiers where you allow it. Anything above your acceptable uplift, or touching a clause you marked always-escalate, stops and waits. You can also require approval on every renewal regardless of terms.
• How accurate is the date extraction? Each extracted obligation is shown with the clause it came from, so you can verify it against the source text rather than trusting a summary. Where an agreement is ambiguous about its own notice mechanics, it is flagged rather than guessed.
• What if we want to let an agreement lapse? Lapsing is a first-class outcome. The assistant can prepare and send the non-renewal notice inside the window, against the notice provisions in the agreement itself, and the record shows when it was served.