TSIA recognizes Inbenta Encore for cutting AI cost per interaction by 40%.
For more than 20 years, over 1,000 organizations including BBVA, Deutsche Bank, Samsung, and FamilySearch have trusted Inbenta to deliver enterprise AI that works. Its flagship Encore platform serves financial services, healthcare, travel, and e-commerce with 98%+ accuracy, prevents hallucinations, full auditability, and 850+ integrations that eliminate vendor lock-in.
DALLAS, Oct. 01, 2026 (GLOBE NEWSWIRE) — Inbenta, an enterprise AI platform for customer experience (CX), today announced it has been named a finalist for the 2026 STAR Award for AI Value Optimization by the Technology & Services Industry Association (TSIA), one of TSIA’s highest honors. The recognition centers on Inbenta’s Encore platform. Inbenta will be recognized alongside this year’s STAR Award honorees at TSIA’s Technology & Services World (TSW) 2026, October 19–21 at ARIA Resort & Casino in Las Vegas.
Highlights
- Conversational AI priced per interaction, not per token
- Observable 40% lower cost per interaction than retrieval-augmented generation (RAG)-first architectures, due to a unique knowledge-first design
- Deployment time is now taking just hours to structure a knowledge base that historically took four employees two months by hand
“Right now, most of the contact center AI market is asking CX leaders to choose between spending more and doing less. That’s a choice nobody should have to make. We built Encore on a simple idea: the cheapest token is the one you never have to spend. Every new customer, channel or language should make the system smarter, not more expensive. Being named a TSIA STAR Award finalist means a lot to our team, because TSIA is the organization that benchmarks this industry. They’re seeing what our customers already see. Real AI value comes from how you build it, not from haggling over the price of a token.” – Melissa Solis, CEO, Inbenta
CX Leaders Need AI Value They Can Prove
“When ROI can’t be explained, budgets are questioned, and teams are downsized,” TSIA wrote in The State of Customer Success 2026. Inbenta’s view: the costliest habit in contact center AI is paying a model to compose answers the business already knows.
Encore Does the Expensive Work Once
Encore uses Inbenta’s proprietary Knowledge Engineering to turn enterprise content into expert-approved question-and-answer “intents” that are stored as verified answers, not generated on the fly. It connects to existing content in 30 to 60 minutes, is model-agnostic and is priced per interaction, not per token.
Customers Are Scaling Service Without Scaling Headcount
“We’ve more than doubled the number of people we’re able to serve. What amazed me most is what it didn’t cost.” – Josh Willis, Founder and Director, Mission Regan
“Inbenta’s chatbot now handles on average 91% of chat inquiries on its own, with less than 10% requiring escalation to a live agent.” – Vanessa Day, Manager, Digital Content, Alterra Mountain Company
See Encore at TSW 2026
Learn more about Inbenta Encore or visit Booth #314 at TSIA’s Technology & Services World 2026, October 19–21 in Las Vegas.
About Inbenta
For more than 20 years, over 1,000 organizations including BBVA, Deutsche Bank, Samsung and FamilySearch have trusted Inbenta to deliver enterprise AI that works. Accuracy and auditability are non-negotiable for organizations where customer experience is serious business. Inbenta delivers both through Encore, a flagship AI orchestration platform serving varied industries, including financial services, healthcare, travel and e-commerce. Customers achieve 75% faster deployment, 50% lower overhead and 35% better first-contact resolution, supported by a model-agnostic design that eliminates vendor lock-in.
Media Contact
Carly Telpner | carly.telpner@inbenta.com
Frequently Asked Questions
What is AI value optimization?
AI value optimization is the practice of using AI to lower cost-to-serve while protecting or growing the revenue each customer relationship produces. TSIA’s STAR Award category of the same name recognizes teams that balance cost-to-serve with high-margin retention and expansion revenue. In a contact center, that means tracking cost per interaction, knowing which interactions AI handles well, and checking whether customers served by AI renew and buy more. The question shifts from “how much AI can we deploy?” to “which AI spend pays for itself?”
How do I lower cost per interaction as we expand AI in customer service?
To lower cost per interaction as AI expands, cut two costs: the tokens each conversation consumes, and the agent time spent when AI fails and escalates. Start with high-volume, repeat questions and answer them from approved content, so you aren’t paying a large language model (LLM) to write each response. Cap the context sent to the model, so each answer uses only the information it needs. Review escalated conversations and fill the content gaps behind them. Track cost per interaction across AI and agents, because an escalation costs both the tokens and the agent’s time.
How much should AI cost per customer interaction?
AI should cost well below a live-agent-handled contact once you count every conversation, including the ones AI fails to resolve. No single benchmark fits every contact center, because factors like prompt size, channel and escalation rate can all move the number. To calculate it, add what AI spends on every conversation to the agent cost of the ones it escalates, then divide by total contacts to get your true cost per interaction. Compare the result with your fully loaded cost per agent-handled contact, including salary, benefits, training and tooling. Pilots tend to handle simple questions, while production brings harder ones that escalate more often, so expect the number to change after launch. Review it monthly by channel and use case, so you can see which ones pay back and which need fixing.
How do I cut AI costs without hurting the customer experience?
Cut what each resolution costs, not the service customers get. Savings that hurt CX tend to come from rationing: capping conversations, hiding the path to an agent to boost containment, or switching to a cheaper model that gets more answers wrong. Savings that protect CX come from changing how answers are produced. Answers grounded in approved content cost less to deliver and stay consistent, because the model isn’t improvising. Passing the full conversation to the agent at handoff shortens handle time and keeps customers from having to start over. Measure customer satisfaction and repeat contacts within seven days alongside cost, so a cheaper month doesn’t hide a worse experience.
A video accompanying this release is available at https://www.globenewswire.com/NewsRoom/AttachmentNg/ff90873c-acd8-4e6f-bcb3-e88f507cc62a
Nguồn: GLOBENEWSWIRE – Đơn vị phát hành hoàn toàn chịu trách nhiệm về nội dung thông báo này.
