What you record for each need
How well your current platform does it: 1 critical gap, 2 major gap, 3 adequate, 4 strong, 5 excellent, or don't know. How you know: seen in our production or vendor says so. Whether it matters for the next contract term: must-have, nice-to-have, not needed. A need marked not needed drops out of every outcome. Nothing is averaged: a must-have gap is a gap whatever the other ratings are, and a rating you do not know is a request for proof, never a low score.
Rules and bands
| Finding | Raised when | Severity |
|---|
| Must-have gap | A must-have is rated 2 or below. | High |
| Proof needed | A must-have is rated 3 or above on the vendor's word, or its rating is not known. | Medium |
| Nice-to-have gap | A nice-to-have is rated 2 or below. It changes no outcome. | Note |
| Layer needs a specialist | Half or more of a non-core layer's must-haves are gaps. Heuristic share. | High |
| Core layer needs a market test | Half or more of a core layer's must-haves are gaps. Heuristic share. | Critical |
| Too many layers need specialists | 3 or more layers need a specialist. Heuristic count. | Critical |
| No time to evaluate | The gate is Run an evaluation and fewer than 6 months remain to the notice date. Heuristic months. | Critical |
| Little time to negotiate | The gate needs conditions and fewer than 3 months remain to the notice date. Heuristic months. | High |
| Exit and data terms unknown | The exit, data-export and transition terms of the current contract are not known. | High |
| Long term with open conditions | The term offered is 3 years or more while the gate needs conditions or an evaluation. Heuristic years. | Medium |
| Renewal clock not entered | Months to the notice date are not entered, so timing is not checked. | Note |
Layer outcomes
| Outcome | When |
|---|
| Renew as is | No must-have is a gap and none needs proof. |
| Renew with conditions | Some must-haves need proof, or fewer than half of the layer's must-haves are gaps. Write the proof, the fix, or a dated roadmap commitment with a remedy if it slips, into the renewal. |
| Add a specialist | Half or more of the layer's must-haves are gaps, on a layer a specialist product can serve beside the platform. |
| Test the market | Half or more of the layer's must-haves are gaps, on a core layer of the platform itself. |
A core layer is what the contact center platform itself is: Routing + Orchestration and Conversation Management. A gap there is not closed by adding a product beside the platform, so it calls for a market test; on the other layers a specialist can serve beside it.
The renewal gate
| Gate | When |
|---|
| Renew | Every layer can renew as is. |
| Renew with conditions | At least one layer needs conditions or a specialist, and no evaluation trigger fires. |
| Run an evaluation | A core layer needs a market test, or 3 or more layers need a specialist. |
The clock then checks whether there is time to act: fewer than 6 months to the notice date is too little to run an evaluation, and fewer than 3 months is little time to negotiate conditions. The next step is Contract Risk Scanner when the exit and data terms are unknown or conditions must be written in, RFP Requirement Builder when the gate says evaluate, and TCO Calculator when the call is to renew.
Thresholds
| Threshold | Value | Basis |
|---|
| Rating at or below which a need is a gap | 2 | Rule of the method |
| Share of a layer's must-haves that are gaps before the layer needs more than a contract fix | 50% | Heuristic, no published source |
| Layers needing a specialist before the stack itself should be evaluated | 3 | Heuristic, no published source |
| Months to run an evaluation before the notice date | 6 | Heuristic, no published source |
| Months to negotiate conditions before the notice date | 3 | Heuristic, no published source |
| Term length, in years, that locks conditions in for too long without a remedy | 3 | Heuristic, no published source |
Layers and needs
L7 Analytics + Governance
- Real-time dashboards you can build yourself, beyond the standard reports
- AI quality assurance (QA) that evaluates every interaction
- Findings from analytics feed back into routing and coaching
- Compliance recording with an audit trail
- Workforce management (WFM) forecasts within 5% at the interval level (usually 15 or 30 minutes)
L6 Routing + Orchestration (core)
- Routing on the customer's intent (what they want), on top of agent skills
- AI handles the contact first and hands over to a person cleanly when needed
- Orchestration across the contact center platform, CRM and WFM
- Queue priority that adjusts to customer value and predicted complexity
- Staffing and routing adjust automatically during the day
L5 Conversation Management (core)
- Voice and digital channels on one platform
- Context carries across channels (a chat that moves to the phone keeps its history)
- Agent desktop brought down to fewer than 3 applications
- Messaging channels (WhatsApp, RCS, Apple Business Chat)
- Asynchronous conversations (the customer replies when they choose) managed to a service level
L4 Reasoning + Planning
- Virtual agent (IVA) built on large language models (LLMs), going beyond fixed intent lists
- Agent assist in real time, during the conversation
- Knowledge AI that answers from your own content (retrieval augmented generation, RAG)
- AI that carries out tasks, such as updating an order, as well as answering
- Guardrails that stop invented answers and policy breaches
L3 Policy + Guardrails
- Card data kept out of PCI scope by tokenization (card numbers never touch your systems)
- AI guardrails for every customer-facing decision
- Identity verification consistent across all channels
- Bias detection and fairness monitoring for AI
- Audit trails for automated decisions
L2 Workflow Execution
- Your top 10 workflows automated from start to finish
- Workflows that other systems can start through an API, as well as from the screen
- Exceptions handled on the contact, without a callback
- Updates written back to other systems, as well as read from them
- Workflow versions, with the ability to roll back
L1 Data Access
- Agent desktop and virtual agent read and update the CRM in real time
- Interaction events streamed to other systems as they happen
- Clean customer data: no duplicates, current and complete
- API access to billing, case management and ERP (enterprise resource planning)
- Data governance with documented ownership
What this tool cannot tell you
- It records how you see your current platform. Nothing here tests the platform, and two people in the same operation can rate it differently. A rating you have seen in production carries more weight than one on the vendor's word, which is why the tool asks how you know.
- The layers and needs describe a modern contact center stack. Use them as a checklist: mark anything you do not need as not needed, and it drops out of every outcome.
- The share, count and month thresholds are heuristics with no published source, and are labelled that way. Set your own if your procurement cycle differs.
- It never names or ranks a vendor. A market test starts from your requirements, which the RFP Builder turns into a structured request.