Menteia · SAVANT AI · October 2026 · Confidential
# Competition: customer complaints and features
**SAVANT AI** — AI contact center on voice, WhatsApp and email that resolves (not just answers) mobile carriers' customer requests.
| Feature | Status | Priority | Phase | Note |
| --- | --- | --- | --- | --- |
| **Agent test bench and regression before publishing** | Missing | High | Months 1–6 | Per-brand test conversations run on every prompt, flow or model change; mitigates inconsistent answers and post-update breakage. |
| **Voice and telephony redundancy (secondary provider and failover)** | Missing | High | Months 7–12 | Currently depends on Vapi and Twilio; roadmap already plans reducing Twilio dependency in months 7–12. Include latency and uptime SLAs. |
| **Public status page and contractual SLA** | Missing | Medium | Months 1–6 | Not present in the product; outages are a recurring complaint for Five9 and voice platforms. |
| **Security certifications (SOC 2 / ISO 27001) and documented LFPDPPP compliance** | Missing | Medium | Months 7–12 | Table stakes for large carrier and BPO procurement; not present in the product. |
| **Standard CRM/helpdesk connectors (Salesforce, HubSpot, Zendesk)** | Missing | Medium | Months 13–18 | Uses its own Frappe Helpdesk today; global competitors ship these connectors out of the box. |
| **Sales and retention module (abandon recovery, anti-port-out offers)** | Missing | Medium | Months 13–18 | Already on the roadmap (months 13–18). |
| **Public API docs and SDK for integrators** | Missing | Low | Months 19–24 | A Botmaker complaint; useful once MVNO enablers and new verticals arrive. |
| **Transparent pricing with cap and cost simulator** | Partial | High | Months 1–6 | Per-resolution and per-minute rates exist, but no public rate card or guaranteed monthly cap; turn the market's #1 complaint into a selling point. |
| **Answer quality control (anti-hallucination and 100% conversation QA)** | Partial | High | Months 1–6 | Code guards exist for critical actions; missing automatic scoring of every answer against the knowledge base and alerts on doubtful replies. |
| **Intent and contact-reason analytics with export** | Partial | Medium | Months 7–12 | Panel has analytics and FCR; missing breakdown by intent, root cause and export/BI for the carrier. |
| **Customer self-service editing of knowledge base and flows** | Partial | Medium | Months 7–12 | Knowledge base is per brand, but there is no evidence the carrier can edit it without Menteia. |
| **Real actions in the carrier's system (balance, porting, suspension, top-ups)** | We have it | — | — | Product core; addresses the 'bots only answer' complaint. |
| **Human handoff with mandatory summary** | We have it | — | — | Twilio Flex with summary and after-hours ticket numbers; tackles the no-exit loop. |
| **Omnichannel voice + WhatsApp + email in one panel** | We have it | — | — | Beats the single-channel focus of Treble/Yalo. |
| **Measured resolution (48 h FCR with survey)** | We have it | — | — | Auditable basis that answers 'what counts as a resolution'. |
| **Brand onboarding in days** | We have it | — | — | Per-brand isolation; contrasts with long Genesys/Five9 rollouts. |
## To the roadmap
### Months 1–6
- **Transparent pricing with cap and cost simulator** (partial, high): Per-resolution and per-minute rates exist, but no public rate card or guaranteed monthly cap; turn the market's #1 complaint into a selling point.
- **Answer quality control (anti-hallucination and 100% conversation QA)** (partial, high): Code guards exist for critical actions; missing automatic scoring of every answer against the knowledge base and alerts on doubtful replies.
- **Agent test bench and regression before publishing** (missing, high): Per-brand test conversations run on every prompt, flow or model change; mitigates inconsistent answers and post-update breakage.
- **Public status page and contractual SLA** (missing, medium): Not present in the product; outages are a recurring complaint for Five9 and voice platforms.
### Months 7–12
- **Voice and telephony redundancy (secondary provider and failover)** (missing, high): Currently depends on Vapi and Twilio; roadmap already plans reducing Twilio dependency in months 7–12. Include latency and uptime SLAs.
- **Intent and contact-reason analytics with export** (partial, medium): Panel has analytics and FCR; missing breakdown by intent, root cause and export/BI for the carrier.
- **Customer self-service editing of knowledge base and flows** (partial, medium): Knowledge base is per brand, but there is no evidence the carrier can edit it without Menteia.
- **Security certifications (SOC 2 / ISO 27001) and documented LFPDPPP compliance** (missing, medium): Table stakes for large carrier and BPO procurement; not present in the product.
### Months 13–18
- **Standard CRM/helpdesk connectors (Salesforce, HubSpot, Zendesk)** (missing, medium): Uses its own Frappe Helpdesk today; global competitors ship these connectors out of the box.
- **Sales and retention module (abandon recovery, anti-port-out offers)** (missing, medium): Already on the roadmap (months 13–18).
### Months 19–24
- **Public API docs and SDK for integrators** (missing, low): A Botmaker complaint; useful once MVNO enablers and new verticals arrive.
## Conclusion
The market complains about bots that answer but don't resolve, loops with no exit to a human, and opaque per-resolution pricing; SAVANT tackles all three head-on with real actions in the telco system, summarized handoff and auditable FCR. Its biggest gap is operational robustness: agent QA and regression testing, voice/telephony redundancy and public SLAs, exactly what a large carrier will demand before signing.