A communications-platform shortlist that runs on demos almost always buys the demo, not the platform. The demo is a script; the evaluation needs a rubric — the same ten sections, scored the same way, for every vendor on the list. This post writes that rubric down once, in a form you can paste into a request-for-proposal document, and it answers each section for Devotel Orbit with a link to the public feature, compare, or docs page a reviewer can verify instead of transcribing the vendor's word.
The rubric is ordered the way a real evaluation runs: channels first, because a vendor that fails coverage fails everything; then the AI agent and contact-center surfaces; then the data layer; then the controls your compliance team owns; then the money questions — billing model, network ownership, pricing transparency; and finally the two sections that kill vendors late, trial access and support tiers. Score each section 1–5, weight the ones your team named as blockers, and the shortlist usually cuts itself from five to two in an afternoon.
How to score any communications platform
- Channel coverage. List every channel you will touch in the next 24 months — SMS, MMS, WhatsApp, RCS, email, voice, video, fax — and check each against one account on the demo. On Orbit, the full channel set ships on one account and one wallet, documented on the CPaaS feature surface, and the channel fallback matrix guide walks the cross-channel routing checks a demo should run.
- AI-agent surface. Ask whether the agent runs on the same platform that carries the calls, or a bolt-on layer handshaking through a separate vendor. Test on the same demo account: build an agent, have it answer a live call, and read the per-action cost. Orbit's AI agent layer is native to the call and message path, with cost governor controls, on the AI-as-a-Service surface; the AI agent failure catalog is the checklist of failure modes to probe.
- Contact-center surface. If you will ever run a support or sales floor, test routing, quality scoring, and the agent desktop now — not after the migration. On Orbit the contact center ships built-in — queues, omnichannel routing, LLM-judged QA scorecards, and the shared inbox — on the CCaaS feature surface; the contact-center software comparison guide is the section-by-section comparison framework.
- CDP surface. Ask where customer identity lives: a native customer data platform under the channels, or a CRM connector you operate yourself. On Orbit the CDP is native — one customer profile, consent record, and conversation history across every channel — on the CDP feature surface; the CDP vs CRM data-layer post frames the question.
- Compliance controls. The only acceptable answer to "whose job is the regulation?" is "yours — here are the controls." Every control must be tenant-owned: configurable by you, defaulted open, and never a platform mandate you cannot audit. Orbit's posture is exactly that — opt-outs, quiet-hours windows, data residency pins, and evidence exports are all tenant-owned surfaces under the compliance docs; the quarterly compliance posture review publishes the control list as evidence.
- Billing model. Per-seat licences punish growth; usage pricing without a public rate card punishes planning. Ask for the arithmetic of your next-24-months usage in the demo, and check which model the vendor is quoting. Orbit bills pay-as-you-go against one wallet with no per-seat licence — the pricing comparison post walks the two models in numbers.
- Network ownership. For outbound voice and messaging, ask one question: whose network physically terminates the traffic? A vendor running its own carrier infrastructure owns routing quality and accountability; a reseller leases both. Orbit terminates outbound traffic over Devotel's own wholesale softswitch as carrier of record, routing across hundreds of interconnected carriers, documented on the carrier of record page and the network ownership section of the enterprise alternatives guide.
- Pricing transparency. A public rate card is a claim you can test; "custom quote" is not. Before the demo, write down three usage numbers; on the demo, compute the bill from the vendor's published page. Orbit publishes per-country and per-channel rates on the pricing page, with no platform fee and no minimums.
- Trial access. A vendor that cannot hand you a test key today is telling you the trial is a sales cycle. Self-serve signup says more about a platform than the demo does. Orbit's self-serve sandbox and live API keys are documented in the quickstart — you can score section 9 from that page alone.
- Support tiers. Ask what is actually sold: included tiers versus paid plans, and what the response-time contract says. And check the public status page before you ask — a vendor without one is asking you to trust support you cannot see.
The scoring sheet
Copy this table into your RFP document; one row per section, one score per vendor, and a reference column that forces every answer to cite a public page.
| # | Section | What to test on the demo | Orbit reference |
|---|---|---|---|
| 1 | Channel coverage | Every planned channel on one account, one wallet | CPaaS feature surface |
| 2 | AI agents | Native agents on the same account, per-action cost visible | AI agent surface |
| 3 | Contact center | Routing, QA scoring, and agent desktop now, not later | CCaaS feature surface |
| 4 | Customer data | Native CDP under the channels, or a connector you operate | CDP feature surface |
| 5 | Compliance controls | Tenant-owned controls, defaulted open, exportable evidence | Compliance docs |
| 6 | Billing model | Work the next-24-months arithmetic in the demo | Per-seat vs usage pricing |
| 7 | Network ownership | Whose network terminates your outbound traffic | Carrier of record |
| 8 | Pricing transparency | Compute the bill from a public rate card | Pricing page |
| 9 | Trial access | A test key today, not a sales cycle | Quickstart |
| 10 | Support tiers | Included tiers versus paid plans, and a public status page | Status page |
A completed row is a vendor answer that cites its own public page. An empty row is a finding. The sheet works on any vendor on the shortlist; the Orbit column exists so a reviewer can verify the answer instead of transcribing it, which is exactly what an RFP scoring sheet should force.
Where the checklist sits in the alternatives hub
Ten sections tell you what to ask any vendor. When you have the answers, the alternatives hub sorts the vendors by business size — small business, growing teams, enterprise — and the per-competitor head-to-head comparisons weighted by the same sections. Run the checklist first; shop by segment second; the two steps are deliberately separate so a strong demo cannot skip section 7 or 8.
Frequently asked questions
Is this checklist for CPaaS or CCaaS evaluations?
Both, deliberately. The ten sections apply to any vendor selling a communications platform — channel coverage, AI agents, contact center, data layer, controls, billing, network, pricing, trial, support. A pure CCaaS shortlist scores sections 3 and 5 double-weight; a pure CPaaS shortlist double-weights sections 1, 7, and 8.
How do I weight the sections?
Weight the sections your team named as blockers in the kickoff. Most teams find one of two patterns: engineering-led teams weight trial access and pricing transparency, because a vendor that fails section 9 never reaches section 1; operations-led teams weight network ownership and the contact-center surface, because section 7 decides delivery quality and section 3 decides whether the migration ever happens.
What does "tenant-owned compliance" mean in an RFP?
It means every compliance control the vendor claims is a surface your team configures, can turn on and off, and can export evidence about — not a platform-side mandate and not a marketing claim. Score the section on the control's existence and auditability, not on the vendor's certifications.
Should the vendor demo drive the score?
No. The demo drives the test; the score comes from the public page the vendor cites when answering. A section scored on what a salesperson said is a finding, not an answer — which is why the reference column in the scoring sheet above is the load-bearing column.
How does this checklist relate to the /compare head-to-head pages?
It is the generic rubric the head-to-head pages assume. The alternatives hub groups the comparisons by business size; this checklist is the scoring method you apply to any of them, and the one a buyer can reuse when a vendor without a head-to-head page enters the shortlist.