NLP-enhanced CRM
Your CRM system augments your supporters, sales and tech staff
capacities well. However, there's much routine in their work, some
frequently asked questions or common dialogues are still handled
manually, often by using predefined text templates.
This situation is a waste of time for your employees and because
of predefined texts it contains pitfalls such as answering a
specific question with a generic text which may not fit as
answer. Imagine how such a "misunderstood" customer must feel.
Reopening of cases and escalation are often the consequences,
accounting for further workload.
An integrated PetaMem NLP/NLU system changes this situation
dramatically. The need for dispatching capacities is eliminated
completedly as language recognition and text categorization reach
a success ratio of 98% and more. Even well trained humans
dispatchers have difficulties to achive that.
A throughput of up to 70 e-mails per minute¹ in dispatching
mode - 24h a day - is sufficient for even large support and sales
centres. Higher throughput is possible with the apropriate hardware.
But state-of-the-art dispatching and spam-filtering isn't where
our systems stop. For specific fields like tech-support and pre-sales
you can maintain completedly artificial subsystems that receive,
process and answer incoming correspondence. These subsystems can be
integrated and act completedly like regular staff.
Depending on the ontology you provide to these systems and the
resulting complexity of rules and inference, an throughput between
20 and 180 emails/h can be achived on standard hardware. For
higher throughput the system is scalable and runs on clusters too.
¹ On a standard two-way Linux
server.
| 
