dtelecom nails 94% intake capture

An insurance-industry user said on May 22 that a trial of dtelecom for claims intake calls captured all required fields during live conversations and reached a 94% first-pass completion rate, according to a post on X. The claim points to one of the most closely watched insurance workflows for voice AI: first notice of loss, or FNOL, where carriers gather the initial facts needed to open and route a claim. dTelecom’s website says it provides real-time voice, speech-to-text and text-to-speech infrastructure for AI agents and developers. ### Why does a single 94% number matter in claims intake? A 94% first-pass completion rate matters because FNOL is the point where insurers collect the core facts needed to start a file, assign the claim and decide whether human follow-up is needed. Industry materials on AI claims intake describe the process as time-sensitive and data-intensive, with delays often caused by missing fields, rekeying and inconsistent handoffs between intake and adjusters. (x.com) Claims-intake vendors and consultants describe the same operational problem in similar terms: if the first call does not capture required information cleanly, staff have to call back, re-enter data or pause downstream work. FurtherAI said legacy intake models create rekeying risk and delay first contact, while Selectsys said AI FNOL tools are designed to flag missing or vague information for follow-up. (x.com) ### What exactly was reported about dtelecom? The May 22 X post said the user had tried dtelecom for claims intake calls and that the AI “captured all required fields” during live calls, with a “first-pass completion rate” of 94%. Reuters could not independently verify the trial design, sample size, line of business, or whether the result covered production traffic or a limited pilot. (furtherai.com) dTelecom’s public site does not market a packaged insurance claims product on the page reviewed, but it does say developers can build real-time voice, video, speech-to-text and text-to-speech applications for humans and autonomous agents. The company also says its network is designed for low-latency communications and pay-per-use access. ### Where would this show up inside an insurer’s workflow? (x.com) FNOL is the stage where callers report an auto accident, property loss or other insured event and the insurer gathers policy, contact, incident and loss details. Industry descriptions of AI FNOL systems say the tools are used to identify claim intent, guide callers through structured intake, validate fields and trigger routing into downstream workflows. (dtelecom.org) A higher first-pass capture rate would typically reduce the number of files that need immediate manual cleanup before triage. Claims-intake specialists say that cleaner intake supports faster handoffs to adjusters, while AI claims-processing vendors say structured intake data can feed fraud review, severity prediction and document workflows. (infinitewatch.ai) ### What can and can’t be concluded from the post? The X post provides a user-reported data point, not an audited carrier case study. Without details on call volume, claim complexity, exception handling, escalation rules and integration into a claims platform, the 94% figure should be treated as a trial result described by the user rather than a verified industry benchmark. (agenttech.io) Deloitte said in its claims-transformation research that insurers are using enabling technologies across a larger share of claims work, while still needing human engagement at critical moments. That framing fits the current market: carriers are testing automation in intake, but they still measure success by whether adjusters receive usable files and fewer incomplete handoffs. (x.com) ### What should readers watch next? The next proof point will be whether dtelecom or the user publishes a fuller case study with call counts, required-field definitions, exception rates and downstream outcomes such as triage speed or reduced rework. dTelecom’s website says its platform supports AI voice, STT and TTS workloads, which would be the components used in a broader insurance deployment. (dtelecom.org) (deloitte.com)