Product: voice
Voice AI that resolves.
Ontological Voice Agents are trained on serviceMob's Experience Data Ontology for your business, so they are grounded in what resolves your real interaction patterns. Every conversation feeds AMPRx, CPRx, 1CX Rate and PECs back into the model.
-
$5.8M
found in repeat-contact reduction alone at one logistics enterprise
-
~15%
what generic voice AI deflects, at limited scope
-
4 signals
AMPRx, CPRx, 1CX Rate and PECs, learned from every call
Definition
What Ontological Voice Agents are.
Ontological Voice Agents are voice AI trained directly on the Experience Data Ontology the Brain builds for your enterprise, so each conversation is grounded in what resolves your actual interaction patterns. Deflection percentage is not resolution. The agents are built to resolve, and each conversation returns to the ontology as a learning signal.
The ontological brain learns from product, marketing, sales and service data, so the agents understand the whole business. When they do, they stop deflecting and start resolving.
Questions
The two questions we hear first.
What are Ontological Voice Agents?
Voice AI agents trained on serviceMob's Experience Data Ontology for your business. They are measured by what generic voice AI is never measured by: whether the experience resolved. 1CX Rate, CPRx, AMPRx and PECs from every conversation feed the continuous learning loop.
How are they priced?
By unique tenant, on usage that scales with contact demand, plus a data hosting fee that covers training and continuous improvement. The figures are in your proposal.
Deflection is not resolution
A deflected contact that does not resolve comes back.
-
Problem
A head of service is under pressure to cut contact-center cost with voice AI. Every vendor in the room quotes a deflection rate.
-
Complication
Generic voice AI is trained on generic conversation patterns, not on the business's experience ontology. It gets people off the phone. Effective deflection sits around 15% at limited scope, a low ceiling, and every deflected-but-unresolved issue returns through another channel, angrier and more expensive the second time. Every dead end and loop adds customer effort, and customer effort is what destroys loyalty.
-
Solution
Ontological Voice Agents are trained on the Experience Data Ontology for that enterprise, so they are grounded in what resolves its real interaction patterns. Each conversation feeds AMPRx, CPRx, 1CX Rate and PECs back into the ontology. The agents get better at resolving, and the business learns which experiences it should prevent upstream.
-
Value
One logistics enterprise found $5.8M in repeat-contact reduction alone. Resolution protects loyalty. Deflection quietly destroys it.
Two categories of voice AI
What each kind is trained on, measured by and does when it misses.
| Generic voice AI | Ontological Voice Agents | |
|---|---|---|
| Trained on | Generic conversation patterns | Your Experience Data Ontology |
| Measured by | Deflection rate | 1CX Rate, CPRx, AMPRx and PECs |
| When it misses | The contact returns through another channel | The miss becomes a learning signal in the ontology |
| What the business learns | How many calls were kept off the phone | Which experiences to prevent upstream |
How we deliver
The people ship with the software.
Ontological Voice Agents arrive as Software with a Service: the agents plus a Forward Deployed team that builds your ontology, trains the agents on it and stays as your business changes. Pricing is set in your proposal.
FDE
Forward Deployed Engineers
Build and validate the data ontology, normalize fragmented data and establish secure pipelines.
FDS
Forward Deployed Strategists
Advisory, roadmaps, operating model design and the recommendations that move the business units causing demand.
FDDS
Forward Deployed Data Scientists
Predictive models, experiential analytics and the statistical validation behind every number we publish.
SME
Subject Matter Experts
Contact center, CRM, workforce management and industry depth, inside your operation from the first week.
Solve. Don't deflect.
Bring your current deflection numbers and your repeat-contact rate to a working session. We will show the gap between the two and what it costs.