Renewal Interrogation

Renewal Interrogation

Generate a vendor-specific playbook for SaaS contract renewals where agent workloads have been added since the last negotiation. Most procurement teams are negotiating with 2023-era contract knowledge against vendors who have been restructuring agent pricing since mid-2025. This skill closes that asymmetry.

When to Invoke

• User says "renewal-interrogation", "renewal playbook", "SaaS renewal prep", "contract negotiation for agents", "help me prep for [vendor] renewal"
• Before any SaaS renewal where AI features, Copilot integrations, or deployed agents have been added since the last contract
• After completing an agent-system-touch-map to translate meter exposure into negotiating leverage
• When a vendor sends a renewal proposal that includes new AI/agent SKUs

Inputs

1. Vendor name — which vendor's renewal is this? (e.g., Salesforce, Microsoft, ServiceNow)
2. Current contract summary — seat count, current spend, contract end date, existing AI/Copilot add-ons (one paragraph is sufficient)
3. Planned agent usage — what agents are deployed or planned? What operations will they perform?
4. Negotiation priorities — what matters most? (cost predictability / cap protection / portability / audit rights / avoid double-billing / exit clauses)

If any input is missing, ask before generating the playbook.

Phase 1: Vendor Profile

Map the vendor to their known agent-pricing behavior patterns. Use the reference below; for vendors not listed, proceed with the generic playbook and flag as [VENDOR NOT IN REFERENCE — verify current pricing model]:

Vendor Behavior Reference (2026-Q2)

Salesforce (Agentforce / Flex Credits)

• Dodge pattern: will not disclose the credit-to-action conversion ratio. Flex Credits are priced as a bundle; the per-action cost is deliberately opaque.
• Leverage points: "Einstein 1" bundles may already include a credit allocation — verify what's covered before buying add-on credits.
• Watch for: "unlimited agent" language that has a velocity cap buried in the order form.

Microsoft (Copilot Credits / Agent 365)

• Dodge pattern: avoids explaining the overlap between Copilot Credits (consumed per Copilot Studio agent turn) and Agent 365 Credits (consumed per autonomous agent action). You can double-spend.
• Leverage points: E3/E5 licenses may already include Copilot usage allowances — press for the exact credit-to-action conversion table.
• Watch for: Entra Agent ID may require a separate licensing SKU depending on tenant configuration.

ServiceNow (Action Fabric / Assist Currency)

• Dodge pattern: governed-path (Flow Designer / Assist actions) vs raw REST API calls have different metering — reps often present only the governed-path price.
• Leverage points: if your agents use raw REST, switching to governed-path reduces cost AND improves audit trails. Use this as a negotiating chip.
• Watch for: "Action Fabric units" pricing varies by edition tier; press for the per-unit rate at your specific edition.

HubSpot (Breeze Credits / Outcomes)

• Dodge pattern: "outcomes-based" billing sounds efficient but "outcome" is defined vendor-side; challenge the definition in writing.
• Leverage points: Breeze usage within existing Marketing/Sales Hub tiers may be included — verify before purchasing add-on credits.
• Watch for: AI sequence automation counts as a Breeze action per send in some SKU configurations.

Atlassian (Rovo Credits)

• Dodge pattern: Rovo credits consumed by agents vs consumed by human AI-assist features are pooled — high agent usage eats human-facing features.
• Leverage points: press for a split-pool option (agent budget vs human budget) to protect human-facing AI.
• Watch for: Jira/Confluence AI suggestions may be bundled or separately metered depending on tier.

Workday (Flex Credits / Agent System of Record)

• Dodge pattern: "Agent System of Record" framing makes agents sound like a neutral feature; the per-request pricing for write/approve operations is not disclosed upfront.
• Leverage points: HCM vs Finance module agents are often metered separately — consolidate workloads to the lower-rate module where possible.

Zendesk (Automated Resolutions)

• Dodge pattern: "resolution" is defined as ticket closure without human touch; vendors count agent-escalated-then-closed tickets as automated, inflating counts.
• Leverage points: press for a clear definition of "resolution" in the contract; negotiate a cap on resolution-count charges per billing period.

SAP (BTP Credits / API Call Units)

• Dodge pattern: governed-path discounts (via SAP Integration Suite) are real but not automatically applied — you must use blessed APIs to qualify.
• Leverage points: legacy ABAP RFC calls do not qualify for the governed-path discount; migrating to SAP Integration Suite APIs reduces cost and is a concession you can extract.

Phase 2: Generate the Playbook

Produce a four-part playbook structured as:

Part 1 — Opening Anchors (first 10 minutes)

Three questions to ask at the opening of the renewal call to establish information asymmetry in your favor:

[Q1] "Before we review the proposal, can you walk me through how [vendor] meters agent operations — specifically the conversion from [credit unit] to individual actions, and whether that ratio is fixed for the contract term?"

[Q2] "Our usage has changed since the last renewal — we've deployed [N] agents that perform [operation types]. Which of those operations consume [meter unit], and at what rate per operation?"

[Q3] "Does this proposal include both the seat cost for our [N] users and a separate metered charge for agent operations those users trigger? If so, we need to see those two buckets itemized."

Customize the bracketed placeholders using the provided vendor name and usage details.

Part 2 — Mid-Conversation Probes

Four targeted probes for the negotiation middle — use after the vendor has presented their numbers:

[P1] Cap protection: "Is there a monthly or annual hard cap on [meter unit] charges? If not, what's the contractual ceiling on variable agent costs?"

[P2] Identity double-billing: "If our agents use delegated user credentials (OAuth/SSO), does that operation incur both a seat charge for the user and a [meter unit] charge for the agent? Show me a sample billing event."

[P3] Governed-path discount: "Does [vendor] offer a lower meter rate for agents that use the governed API path vs raw REST/API calls? If yes, is that discount contractually guaranteed?"

[P4] Audit rights: "What usage reporting is available — can we pull per-operation meter consumption by agent identity, with timestamps, via API or downloadable report? What's the SLA on billing dispute resolution?"

Part 3 — Closing Protection Clauses

Three clauses to request in writing before signing. These are not standard inclusions — you must ask:

[C1] Meter-change notice: "Any change to the [meter unit]-to-action conversion ratio or meter definition requires 90 days written notice and gives us the right to renegotiate pricing without penalty."

[C2] Volumetric cap: "Agent-generated [meter unit] charges are capped at $[N]/month. Overages require pre-approval before billing."

[C3] Exit portability: "On contract termination, we retain the right to export all agent configuration, workflow definitions, and historical meter usage data in a standard format within 30 days, at no additional charge."

Part 4 — The One Question They Will Try to Dodge

Surface the vendor-specific dodge question — the question whose answer would most shift the negotiation:

Generate this from the vendor profile (Phase 1). Format as:

## The One Question [Vendor] Will Try to Dodge

**The question:** "[exact question text]"

**Why they dodge it:** [one sentence on what the answer reveals]

**How to press:** If they redirect or say "that depends on your configuration," respond: "Can you put the answer in writing in the order form? If the rate is genuinely configuration-dependent, we need the min/max range and the conditions that determine it documented before we sign."

**What a good answer looks like:** [one sentence — what transparency looks like on this specific point]

Rules

• Every output must be specific to the vendor named. Generic "ask about pricing" advice is not acceptable.
• If the user provides contract terms, surface specific gaps — don't produce a generic playbook when vendor-specific leverage is available.
• Closing protection clauses must be framed as specific contract language, not as suggestions to "ask about" something.
• The dodge question must name the specific opacity tactic for that vendor — not a generic "ask about hidden fees."
• If the vendor is not in the reference table, produce the generic playbook and clearly flag it as unverified for that specific vendor.

Verification Checklist

• All four playbook sections present (opening anchors, probes, closing clauses, dodge question)
• Opening anchors customized with vendor name, meter unit, and user's operation types
• Closing clauses written as specific contract language, not suggestions
• Dodge question names the vendor-specific opacity tactic
• Vendor-not-in-reference case flagged if applicable

Source

Nate's Newsletter, 2026-05-15 — "SaaS Agent Licensing: What Your 2026 Renewal Will Look Like" Core insight: SaaS vendors are rewriting 2026 contracts around an agent meter procurement teams don't yet know how to negotiate. A vendor-specific interrogation playbook that distinguishes fair agent licensing (transparent, capped, portable, identity-aware) from rent-seeking (opaque, uncapped, lock-in, double-bills human + agent) is the procurement team's primary defense.