India \u092e\u0947\u0902 business mostly phone calls \u092a\u0930 \u091a\u0932\u0924\u093e \u0939\u0948, \u0914\u0930 \u0935\u0939 calls \u0905\u0915\u094d\u0938\u0930 \u090f\u0915 \u0939\u0940 sentence \u092e\u0947\u0902 Hindi, English \u0914\u0930 \u090f\u0915 regional language \u0915\u094b \u092e\u093f\u0932\u093e\u0915\u0930 \u0939\u094b\u0924\u0940 \u0939\u0948\u0902\u0964 \u0910\u0938\u0940 code-mixed \u092c\u093e\u0924\u091a\u0940\u0924 \u0915\u094b \u0938\u0939\u0940 \u0924\u094c\u0930 \u092a\u0930 text \u092e\u0947\u0902 \u092c\u0926\u0932\u0928\u093e \u092c\u0939\u0941\u0924 \u092e\u0941\u0936\u094d\u0915\u093f\u0932 \u0915\u093e\u092e \u0939\u0948\u0964 Bengaluru \u0915\u0940 company Gnani AI \u0928\u0947 28 October 2026 \u0915\u094b \u0907\u0938\u0940 problem \u0915\u0947 \u0932\u093f\u090f \u0905\u092a\u0928\u093e \u0928\u092f\u093e model launch \u0915\u093f\u092f\u093e \u0939\u0948, \u091c\u093f\u0938\u0915\u093e \u0928\u093e\u092e \u0939\u0948 Prisma v3.0\u0964
Prisma v3.0 \u090f\u0915 speech to text model \u0939\u0948, \u092e\u0924\u0932\u092c \u091c\u094b \u092c\u094b\u0932\u093e \u0917\u092f\u093e \u0939\u0948 \u0909\u0938\u0947 \u0932\u093f\u0916\u0947 \u0939\u0941\u090f text \u092e\u0947\u0902 \u092c\u0926\u0932\u0928\u0947 \u0935\u093e\u0932\u093e system\u0964 \u092f\u0939 model \u0916\u093e\u0938 \u0924\u094c\u0930 \u092a\u0930 Indian telephone audio \u0914\u0930 multilingual conversations \u0915\u0947 \u0932\u093f\u090f train \u0915\u093f\u092f\u093e \u0917\u092f\u093e \u0939\u0948\u0964 \u0907\u0938\u0915\u093e \u092e\u0924\u0932\u092c \u0939\u0948 \u0915\u093f \u092f\u0939 \u0909\u0928 calls \u0915\u094b \u0938\u092e\u091d\u0928\u0947 \u0915\u0947 \u0932\u093f\u090f \u092c\u0928\u093e \u0939\u0948 \u091c\u094b \u0906\u092e \u0924\u094c\u0930 \u092a\u0930 Indian offices, banks \u0914\u0930 customer support \u092e\u0947\u0902 \u0939\u094b\u0924\u0940 \u0939\u0948\u0902\u0964<\/p>\n\n\n\n
Prisma v3.0 Gnani AI \u0915\u093e latest speech to text model \u0939\u0948\u0964 \u0907\u0938\u0915\u093e \u0938\u092c\u0938\u0947 \u092c\u0921\u093c\u093e focus Indian conversations \u0939\u0948\u0902\u0964 \u0906\u092e speech tools mostly \u0938\u093e\u092b-\u0938\u093e\u092b \u092c\u094b\u0932\u0940 \u0917\u0908 English \u092a\u0930 \u0905\u091a\u094d\u091b\u093e \u0915\u093e\u092e \u0915\u0930\u0924\u0947 \u0939\u0948\u0902, \u0932\u0947\u0915\u093f\u0928 \u091c\u092c call \u092e\u0947\u0902 \u0936\u094b\u0930 \u0939\u094b, phone \u0915\u0940 quality \u0916\u0930\u093e\u092c \u0939\u094b, \u092f\u093e speaker \u092c\u093e\u0930-\u092c\u093e\u0930 language \u092c\u0926\u0932 \u0930\u0939\u093e \u0939\u094b, \u0924\u094b \u0909\u0928\u0915\u0940 accuracy \u0917\u093f\u0930 \u091c\u093e\u0924\u0940 \u0939\u0948\u0964
Gnani AI \u0915\u0947 \u092e\u0941\u0924\u093e\u092c\u093f\u0915 Prisma v3.0 \u0907\u0938\u0940 gap \u0915\u094b \u092d\u0930\u0928\u0947 \u0915\u0947 \u0932\u093f\u090f \u092c\u0928\u093e \u0939\u0948\u0964 \u092f\u0939 \u090f\u0915 \u0939\u0940 system \u0938\u0947 9 Indian languages \u0914\u0930 Indian English \u0915\u094b \u0938\u0902\u092d\u093e\u0932\u0924\u093e \u0939\u0948\u0964 \u0907\u0938\u0915\u093e \u092e\u0924\u0932\u092c \u0939\u0948 \u0915\u093f \u0905\u0932\u0917-\u0905\u0932\u0917 language \u0915\u0947 \u0932\u093f\u090f \u0905\u0932\u0917-\u0905\u0932\u0917 model \u0932\u0917\u093e\u0928\u0947 \u0915\u0940 \u091c\u0930\u0942\u0930\u0924 \u0928\u0939\u0940\u0902 \u0939\u0948\u0964 \u090f\u0915 \u0939\u0940 model \u092a\u0942\u0930\u0940 conversation \u0915\u094b \u0938\u092e\u091d \u0938\u0915\u0924\u093e \u0939\u0948, \u091a\u093e\u0939\u0947 \u0909\u0938\u092e\u0947\u0902 language \u092c\u0940\u091a \u092e\u0947\u0902 \u0939\u0940 \u0915\u094d\u092f\u094b\u0902 \u0928\u093e \u092c\u0926\u0932 \u091c\u093e\u090f\u0964
\u092f\u0939 model native 8 kHz telephone audio \u0915\u094b \u092d\u0940 accept \u0915\u0930\u0924\u093e \u0939\u0948\u0964 8 kHz \u0935\u0939 lower bandwidth format \u0939\u0948 \u091c\u094b traditional telephone systems \u092e\u0947\u0902 use \u0939\u094b\u0924\u093e \u0939\u0948\u0964 \u092c\u0939\u0941\u0924 \u0938\u0947 \u092a\u0941\u0930\u093e\u0928\u0947 call recording systems \u0907\u0938\u0940 format \u092e\u0947\u0902 audio save \u0915\u0930\u0924\u0947 \u0939\u0948\u0902, \u0907\u0938\u0932\u093f\u090f \u0907\u0938 support \u0915\u093e \u092e\u0924\u0932\u092c \u0939\u0948 \u0915\u093f companies \u0915\u094b \u0905\u092a\u0928\u0940 \u092a\u0941\u0930\u093e\u0928\u0940 recordings \u0915\u094b convert \u0915\u0930\u0928\u0947 \u0915\u0940 \u091c\u0930\u0942\u0930\u0924 \u0928\u0939\u0940\u0902 \u092a\u0921\u093c\u0947\u0917\u0940\u0964<\/p>\n\n\n\n
Prisma v3.0 \u092e\u0947\u0902 9 Indian languages \u0915\u0947 \u0938\u093e\u0925 Indian English supported \u0939\u0948\u0964 \u092f\u0939 languages \u0939\u0948\u0902 Hindi, Bengali, Marathi, Gujarati, Punjabi, Kannada, Telugu, Tamil \u0914\u0930 Malayalam, \u0914\u0930 \u0938\u093e\u0925 \u092e\u0947\u0902 English\u0964
\u090f\u0915 typical Indian support call \u0915\u093e example \u0932\u0940\u091c\u093f\u090f\u0964 Customer \u0915\u0939\u0924\u093e \u0939\u0948 \u0915\u093f \u0909\u0938\u0915\u093e loan \u0915\u093e EMI deduct \u0928\u0939\u0940\u0902 \u0939\u0941\u0906, \u092b\u093f\u0930 \u0935\u0939 Hindi \u092e\u0947\u0902 \u0905\u092a\u0928\u0940 problem \u0938\u092e\u091d\u093e\u0924\u093e \u0939\u0948, \u0914\u0930 \u092c\u0940\u091a \u092e\u0947\u0902 \u090f\u0915 number Marathi \u092f\u093e Tamil \u092e\u0947\u0902 \u092c\u094b\u0932 \u0926\u0947\u0924\u093e \u0939\u0948\u0964 \u0910\u0938\u0947 mixed sentence \u0915\u094b \u0905\u0932\u0917-\u0905\u0932\u0917 models \u0938\u0947 \u0938\u092e\u091d\u0928\u093e \u092e\u0941\u0936\u094d\u0915\u093f\u0932 \u0939\u094b\u0924\u093e \u0939\u0948, \u0915\u094d\u092f\u094b\u0902\u0915\u093f \u0939\u0930 model \u0938\u093f\u0930\u094d\u092b \u0905\u092a\u0928\u0940 language \u092a\u0930 focus \u0915\u0930\u0924\u093e \u0939\u0948\u0964 Prisma v3.0 \u0915\u093e \u0926\u093e\u0935\u093e \u0939\u0948 \u0915\u093f \u0935\u0939 \u0910\u0938\u0940 mixed \u092c\u093e\u0924 \u0915\u094b \u090f\u0915 \u0939\u0940 flow \u092e\u0947\u0902 \u0938\u092e\u091d \u0938\u0915\u0924\u093e \u0939\u0948, \u092c\u093f\u0928\u093e language \u092c\u0926\u0932\u0928\u0947 \u092a\u0930 \u0905\u091f\u0915\u0947\u0964
\u0907\u0938\u0915\u0947 \u0932\u093f\u090f automatic language identification \u092d\u0940 \u0926\u0940 \u0917\u0908 \u0939\u0948, \u091c\u093f\u0938\u0947 Auto-LID \u0915\u0939\u0924\u0947 \u0939\u0948\u0902\u0964 Company \u0915\u0947 \u092e\u0941\u0924\u093e\u092c\u093f\u0915 3 second \u0938\u0947 \u0932\u0902\u092c\u0940 audio clips \u092a\u0930 \u092f\u0939 feature 98 percent \u0938\u0947 \u091c\u094d\u092f\u093e\u0926\u093e \u092c\u093e\u0930 \u0938\u0939\u0940 language \u092a\u0939\u091a\u093e\u0928 \u0932\u0947\u0924\u093e \u0939\u0948\u0964 \u092f\u0939 company \u0915\u093e \u0905\u092a\u0928\u093e \u0926\u093e\u0935\u093e \u0939\u0948, \u0914\u0930 \u0907\u0938\u0947 independent test \u0938\u0947 verify \u0915\u0930\u0928\u093e \u0905\u092d\u0940 \u092c\u093e\u0915\u0940 \u0939\u0948, \u0932\u0947\u0915\u093f\u0928 \u0905\u0917\u0930 \u092f\u0939 real \u0926\u0941\u0928\u093f\u092f\u093e \u092e\u0947\u0902 \u092d\u0940 \u0907\u0924\u0928\u093e accurate \u0939\u0948, \u0924\u094b \u092f\u0939 mixed calls \u0915\u0947 \u0932\u093f\u090f \u092c\u0939\u0941\u0924 \u092c\u0921\u093c\u093e \u092b\u093e\u092f\u0926\u093e \u0939\u0948\u0964<\/p>\n\n\n\n
Prisma v3.0 \u0915\u0940 \u0938\u092c\u0938\u0947 useful \u092c\u093e\u0924 \u0907\u0938\u0915\u0947 4 transcription modes \u0939\u0948\u0902\u0964 \u0939\u0930 team \u0915\u094b transcript \u090f\u0915 \u0939\u0940 \u0924\u0930\u0940\u0915\u0947 \u0938\u0947 \u0928\u0939\u0940\u0902 \u091a\u093e\u0939\u093f\u090f \u0939\u094b\u0924\u093e\u0964 \u0915\u093f\u0938\u0940 \u0915\u094b audit \u0915\u0947 \u0932\u093f\u090f \u092a\u0942\u0930\u093e record \u091a\u093e\u0939\u093f\u090f, \u0924\u094b \u0915\u093f\u0938\u0940 \u0915\u094b \u092a\u0922\u093c\u0928\u0947 \u0932\u093e\u092f\u0915 clean text \u091a\u093e\u0939\u093f\u090f\u0964 \u0907\u0938\u0932\u093f\u090f Gnani AI \u0928\u0947 \u090f\u0915 \u0939\u0940 model \u092e\u0947\u0902 4 \u0924\u0930\u0940\u0915\u0947 \u0926\u093f\u090f \u0939\u0948\u0902, \u091c\u093f\u0928\u094d\u0939\u0947\u0902 \u0939\u0930 request \u092a\u0930 \u0905\u0932\u0917 \u0938\u0947 \u091a\u0941\u0928\u093e \u091c\u093e \u0938\u0915\u0924\u093e \u0939\u0948\u0964
\u092a\u0939\u0932\u093e mode \u0939\u0948 Verbatim\u0964 \u0907\u0938\u092e\u0947\u0902 \u0939\u0930 filler word \u0914\u0930 repetition \u092d\u0940 \u0932\u093f\u0916\u093e \u091c\u093e\u0924\u093e \u0939\u0948, \u091c\u0948\u0938\u0947 umm, \u0939\u093e\u0902 \u0939\u093e\u0902, \u092f\u093e \u0926\u094b\u092c\u093e\u0930\u093e \u092c\u094b\u0932\u0940 \u0917\u0908 \u092c\u093e\u0924\u0964 \u092f\u0939 compliance \u0914\u0930 legal teams \u0915\u0947 \u0932\u093f\u090f \u091c\u0930\u0942\u0930\u0940 \u0939\u094b\u0924\u093e \u0939\u0948, \u091c\u0939\u093e\u0902 \u092a\u0942\u0930\u093e audit trail \u091a\u093e\u0939\u093f\u090f \u0939\u094b\u0924\u093e \u0939\u0948\u0964
\u0926\u0942\u0938\u0930\u093e mode \u0939\u0948 Transcribe\u0964 \u0907\u0938\u092e\u0947\u0902 clean, \u092a\u0922\u093c\u0928\u0947 \u0932\u093e\u092f\u0915 text \u092e\u093f\u0932\u0924\u093e \u0939\u0948, \u0909\u0938 language \u0915\u0940 native script \u092e\u0947\u0902\u0964 Amounts, dates \u0914\u0930 percentages digits \u092e\u0947\u0902 \u0932\u093f\u0916\u0947 \u091c\u093e\u0924\u0947 \u0939\u0948\u0902, \u0924\u093e\u0915\u093f \u0935\u0939 \u0938\u0940\u0927\u093e business system \u092e\u0947\u0902 use \u0939\u094b \u0938\u0915\u0947\u0902\u0964
\u0924\u0940\u0938\u0930\u093e mode \u0939\u0948 Code-mix\u0964 \u0907\u0938\u092e\u0947\u0902 \u0939\u0930 \u0936\u092c\u094d\u0926 \u0909\u0938\u0940 script \u092e\u0947\u0902 \u0932\u093f\u0916\u093e \u091c\u093e\u0924\u093e \u0939\u0948 \u091c\u093f\u0938 language \u092e\u0947\u0902 \u0935\u0939 \u092c\u094b\u0932\u093e \u0917\u092f\u093e \u0925\u093e, \u0914\u0930 English terms English \u092e\u0947\u0902 \u0939\u0940 \u0930\u0939\u0924\u0947 \u0939\u0948\u0902\u0964 \u091c\u0948\u0938\u0947 \u0905\u0917\u0930 \u0915\u093f\u0938\u0940 \u0928\u0947 Hindi sentence \u092e\u0947\u0902 English \u0915\u093e bank term \u092c\u094b\u0932\u093e, \u0924\u094b \u0935\u0939 term English \u092e\u0947\u0902 \u0939\u0940 \u0932\u093f\u0916\u093e \u091c\u093e\u090f\u0917\u093e\u0964 \u092f\u0939 \u0909\u0928 teams \u0915\u0947 \u0932\u093f\u090f useful \u0939\u0948 \u091c\u094b \u092c\u093e\u0924\u091a\u0940\u0924 \u0915\u094b \u0909\u0938\u0915\u0947 \u0905\u0938\u0932\u0940 \u0930\u0942\u092a \u092e\u0947\u0902 \u0926\u0947\u0916\u0928\u093e \u091a\u093e\u0939\u0924\u0940 \u0939\u0948\u0902\u0964
\u091a\u094c\u0925\u093e mode \u0939\u0948 Transliterate\u0964 \u0907\u0938\u092e\u0947\u0902 \u092a\u0942\u0930\u093e transcript English letters \u092e\u0947\u0902 \u0932\u093f\u0916\u093e \u091c\u093e\u0924\u093e \u0939\u0948, \u091a\u093e\u0939\u0947 \u092c\u093e\u0924 \u0915\u093f\u0938\u0940 \u092d\u0940 language \u092e\u0947\u0902 \u0939\u0941\u0908 \u0939\u094b\u0964 \u092e\u0924\u0932\u092c Hindi \u0915\u0947 \u0936\u092c\u094d\u0926 \u092d\u0940 Roman letters \u092e\u0947\u0902 \u0932\u093f\u0916\u0947 \u091c\u093e\u090f\u0902\u0917\u0947, \u0909\u0928\u0915\u093e translation \u0928\u0939\u0940\u0902 \u0939\u094b\u0917\u093e\u0964 \u092f\u0939 \u0909\u0928 \u0932\u094b\u0917\u094b\u0902 \u0915\u0947 \u0932\u093f\u090f helpful \u0939\u0948 \u091c\u094b Indian languages \u092c\u094b\u0932 \u0924\u094b \u0932\u0947\u0924\u0947 \u0939\u0948\u0902, \u0932\u0947\u0915\u093f\u0928 \u0909\u0928\u0915\u0940 script \u092a\u0922\u093c\u0928\u093e \u092e\u0941\u0936\u094d\u0915\u093f\u0932 \u0932\u0917\u0924\u093e \u0939\u0948\u0964<\/p>\n\n\n\n
Indian business calls \u092e\u0947\u0902 \u0938\u092c\u0938\u0947 \u092c\u0921\u093c\u093e risk numbers \u0915\u093e \u0939\u094b\u0924\u093e \u0939\u0948\u0964 \u0905\u0917\u0930 customer \u0928\u0947 \u0915\u0939\u093e \u0924\u0940\u0928 \u0932\u093e\u0916 \u092a\u091a\u093e\u0938 \u0939\u091c\u093e\u0930, \u0914\u0930 transcript \u092e\u0947\u0902 \u0935\u0939 \u0917\u0932\u0924 \u0932\u093f\u0916\u093e \u0917\u092f\u093e, \u0924\u094b loan \u092f\u093e claim \u0915\u093e \u092a\u0942\u0930\u093e calculation \u092c\u093f\u0917\u0921\u093c \u0938\u0915\u0924\u093e \u0939\u0948\u0964 \u0907\u0938\u0932\u093f\u090f Prisma v3.0 \u092e\u0947\u0902 inverse text normalization \u0926\u0940 \u0917\u0908 \u0939\u0948, \u091c\u093f\u0938\u0947 ITN \u0915\u0939\u0924\u0947 \u0939\u0948\u0902\u0964
ITN \u0915\u093e \u092e\u0924\u0932\u092c \u0939\u0948 \u0915\u093f \u092c\u094b\u0932\u0947 \u0917\u090f numbers \u0915\u094b \u0938\u0939\u0940 digits \u092e\u0947\u0902 \u092c\u0926\u0932 \u0926\u0947\u0928\u093e, Indian digit grouping \u0915\u0947 \u0938\u093e\u0925\u0964 \u091c\u0948\u0938\u0947 \u0924\u0940\u0928 \u0932\u093e\u0916 \u092a\u091a\u093e\u0938 \u0939\u091c\u093e\u0930 \u0915\u094b 3,50,000 \u0932\u093f\u0916\u093e \u091c\u093e\u090f\u0917\u093e, \u0928\u093e \u0915\u093f 350,000\u0964 \u0907\u0938\u0940 \u0924\u0930\u0939 Marathi \u092e\u0947\u0902 \u0938\u093e\u0921\u0947 \u0926\u0939\u093e \u091c\u0948\u0938\u0947 time \u0915\u094b 10:30 \u0932\u093f\u0916\u093e \u091c\u093e\u090f\u0917\u093e\u0964 \u092f\u0939 \u091b\u094b\u091f\u0940 \u092c\u093e\u0924 \u0932\u0917 \u0938\u0915\u0924\u0940 \u0939\u0948, \u0932\u0947\u0915\u093f\u0928 banking, lending \u0914\u0930 claims \u091c\u0948\u0938\u0947 \u0915\u093e\u092e \u092e\u0947\u0902 \u092f\u0939 accuracy \u0939\u0940 \u0938\u092c \u0915\u0941\u091b \u0939\u094b\u0924\u0940 \u0939\u0948\u0964
\u0907\u0938\u0915\u0947 \u0905\u0932\u093e\u0935\u093e developers \u0939\u0930 request \u092e\u0947\u0902 200 \u0924\u0915 custom keywords \u0926\u0947 \u0938\u0915\u0924\u0947 \u0939\u0948\u0902\u0964 \u092e\u0924\u0932\u092c \u0905\u0917\u0930 \u0915\u093f\u0938\u0940 company \u0915\u0947 product \u0915\u0947 \u0928\u093e\u092e \u0905\u091c\u0940\u092c \u0939\u0948\u0902 \u092f\u093e \u0915\u094b\u0908 brand jargon \u0939\u0948, \u0924\u094b \u0935\u0939 keywords \u0926\u0947\u0915\u0930 \u092f\u0939 \u0938\u0941\u0928\u093f\u0936\u094d\u091a\u093f\u0924 \u0915\u0930 \u0938\u0915\u0924\u093e \u0939\u0948 \u0915\u093f model \u0909\u0928\u094d\u0939\u0947\u0902 \u0938\u0939\u0940 \u0932\u093f\u0916\u0947, \u0928\u093e \u0915\u093f \u0915\u093f\u0938\u0940 \u092e\u093f\u0932\u0924\u0947-\u091c\u0941\u0932\u0924\u0947 \u0906\u092e \u0936\u092c\u094d\u0926 \u0938\u0947 \u092c\u0926\u0932 \u0926\u0947\u0964<\/p>\n\n\n\n
Accuracy \u0915\u0947 \u092e\u093e\u092e\u0932\u0947 \u092e\u0947\u0902 Gnani AI \u0928\u0947 Vistaar benchmark \u0915\u0947 results \u0926\u093f\u090f \u0939\u0948\u0902\u0964 Vistaar \u090f\u0915 speech recognition benchmark \u0939\u0948 \u091c\u094b Indian languages \u0915\u0947 \u0932\u093f\u090f use \u0939\u094b\u0924\u093e \u0939\u0948\u0964 Company \u0915\u0947 \u092e\u0941\u0924\u093e\u092c\u093f\u0915 Prisma v3.0 \u0928\u0947 \u0907\u0938 benchmark \u092a\u0930 16.93 percent macro word error rate record \u0915\u093f\u092f\u093e, \u091c\u092c\u0915\u093f ElevenLabs \u0915\u0947 Scribe v2 \u0915\u093e WER 19.98 percent \u0930\u0939\u093e\u0964
Word error rate \u0915\u093e \u092e\u0924\u0932\u092c \u0939\u0948 transcript \u092e\u0947\u0902 \u0917\u0932\u0924 \u0936\u092c\u094d\u0926\u094b\u0902 \u0915\u093e percentage\u0964 \u0915\u092e WER \u0915\u093e \u092e\u0924\u0932\u092c \u0939\u0948 \u091c\u094d\u092f\u093e\u0926\u093e accurate transcript\u0964 Company \u0915\u0947 \u092e\u0941\u0924\u093e\u092c\u093f\u0915 \u092f\u0939 \u0932\u0917\u092d\u0917 15.3 percent relative reduction \u0939\u0948, \u092e\u0924\u0932\u092c \u092a\u0939\u0932\u0947 \u0915\u0947 \u092e\u0941\u0915\u093e\u092c\u0932\u0947 \u0917\u0932\u0924\u093f\u092f\u093e\u0902 \u0907\u0924\u0928\u0940 \u0915\u092e \u0939\u0941\u0908 \u0939\u0948\u0902\u0964
\u092f\u0939\u093e\u0902 \u092f\u0939 \u0938\u092e\u091d\u0928\u093e \u091c\u0930\u0942\u0930\u0940 \u0939\u0948 \u0915\u093f \u092f\u0939 numbers company \u0915\u0940 \u0905\u092a\u0928\u0940 evaluation \u0938\u0947 \u0906\u090f \u0939\u0948\u0902, Vistaar scoring protocol \u092a\u0930\u0964 \u092f\u0939 independent verification \u0928\u0939\u0940\u0902 \u0939\u0948\u0964 \u0907\u0938\u0932\u093f\u090f \u0907\u0928\u094d\u0939\u0947\u0902 \u090f\u0915 promising signal \u0915\u0940 \u0924\u0930\u0939 \u0926\u0947\u0916\u0928\u093e \u091a\u093e\u0939\u093f\u090f, \u0928\u093e \u0915\u093f final proof \u0915\u0940 \u0924\u0930\u0939\u0964 Real \u0926\u0941\u0928\u093f\u092f\u093e \u092e\u0947\u0902 accuracy audio quality, accent \u0914\u0930 background noise \u092a\u0930 \u092d\u0940 depend \u0915\u0930\u0947\u0917\u0940\u0964<\/p>\n\n\n\n
Prisma v3.0 developers \u0915\u0947 \u0932\u093f\u090f company \u0915\u0947 developer API \u0915\u0947 through available \u0939\u0948\u0964 \u0907\u0938\u092e\u0947\u0902 REST, WebSocket streaming \u0914\u0930 batch processing \u0915\u093e support \u0926\u093f\u092f\u093e \u0917\u092f\u093e \u0939\u0948\u0964 REST \u0915\u093e \u092e\u0924\u0932\u092c \u0939\u0948 simple API calls, WebSocket streaming \u0915\u093e \u092e\u0924\u0932\u092c \u0939\u0948 live \u091c\u0948\u0938\u0947 \u091a\u0932\u0924\u0947 \u0939\u0941\u090f audio \u0915\u093e processing, \u0914\u0930 batch \u0915\u093e \u092e\u0924\u0932\u092c \u0939\u0948 \u090f\u0915 \u0938\u093e\u0925 \u0915\u0908 recordings \u0915\u094b process \u0915\u0930\u0928\u093e\u0964
\u092f\u0939 model Gnani \u0915\u0947 agent builder platform \u0915\u0947 through \u092d\u0940 available \u092c\u0924\u093e\u092f\u093e \u0917\u092f\u093e \u0939\u0948\u0964 \u0907\u0938\u0938\u0947 \u0935\u0939 companies \u092d\u0940 \u0907\u0938\u0915\u094b use \u0915\u0930 \u0938\u0915\u0924\u0940 \u0939\u0948\u0902 \u091c\u094b \u0916\u0941\u0926 \u091c\u094d\u092f\u093e\u0926\u093e code \u0928\u0939\u0940\u0902 \u0932\u093f\u0916\u0928\u093e \u091a\u093e\u0939\u0924\u0940, \u092c\u0932\u094d\u0915\u093f \u090f\u0915 ready platform \u092a\u0930 \u0905\u092a\u0928\u093e voice agent \u092c\u0928\u093e\u0928\u093e \u091a\u093e\u0939\u0924\u0940 \u0939\u0948\u0902\u0964
\u0938\u092c\u0938\u0947 \u091c\u094d\u092f\u093e\u0926\u093e \u092b\u093e\u092f\u0926\u093e banks, insurance companies, telecom operators \u0914\u0930 customer support teams \u0915\u094b \u0939\u094b\u0917\u093e, \u091c\u094b \u0930\u094b\u091c \u0939\u091c\u093e\u0930\u094b\u0902 calls handle \u0915\u0930\u0924\u0947 \u0939\u0948\u0902\u0964 \u0907\u0928 calls \u0915\u093e transcript CRM, lending \u0914\u0930 claims system \u092e\u0947\u0902 \u091c\u093e\u0924\u093e \u0939\u0948, \u091c\u0939\u093e\u0902 manual typing \u092e\u0947\u0902 time \u0932\u0917\u0924\u093e \u0939\u0948 \u0914\u0930 \u0917\u0932\u0924\u0940 \u0915\u093e risk \u0930\u0939\u0924\u093e \u0939\u0948\u0964 Gnani AI \u0915\u0947 \u092e\u0941\u0924\u093e\u092c\u093f\u0915 \u0909\u0938\u0915\u0940 speech technology \u092a\u0939\u0932\u0947 \u0938\u0947 \u0939\u0940 \u0930\u094b\u091c 50 million voice interactions \u0915\u094b power \u0915\u0930\u0924\u0940 \u0939\u0948, \u0907\u0938\u0932\u093f\u090f Prisma v3.0 \u0909\u0938\u0940 experience \u092a\u0930 \u092c\u0928\u093e \u0939\u0948\u0964
Price \u0915\u0947 \u092c\u093e\u0930\u0947 \u092e\u0947\u0902 sources \u092e\u0947\u0902 \u0915\u094b\u0908 public price announce \u0928\u0939\u0940\u0902 \u0915\u0940 \u0917\u0908 \u0939\u0948\u0964 \u0907\u0938\u0932\u093f\u090f \u0907\u0938 article \u092e\u0947\u0902 \u0915\u094b\u0908 price claim \u0928\u0939\u0940\u0902 \u0915\u093f\u092f\u093e \u091c\u093e \u0930\u0939\u093e \u0939\u0948\u0964 \u091c\u094b business \u0907\u0938\u0915\u094b use \u0915\u0930\u0928\u093e \u091a\u093e\u0939\u0924\u093e \u0939\u0948, \u0909\u0938\u0947 Gnani AI \u0938\u0947 API access \u0914\u0930 pricing \u0915\u0947 \u092c\u093e\u0930\u0947 \u092e\u0947\u0902 \u0938\u0940\u0927\u093e \u092a\u0942\u091b\u0928\u093e \u091a\u093e\u0939\u093f\u090f\u0964<\/p>\n\n\n\n
Gnani AI \u0928\u0947 Prisma v3.0 \u0915\u094b 28 October 2026 \u0915\u094b launch \u0915\u093f\u092f\u093e\u0964<\/p>\n\n\n\n
\u092f\u0939 9 Indian languages, Hindi, Bengali, Marathi, Gujarati, Punjabi, Kannada, Telugu, Tamil \u0914\u0930 Malayalam, \u0915\u0947 \u0938\u093e\u0925 Indian English \u0915\u094b support \u0915\u0930\u0924\u093e \u0939\u0948\u0964<\/p>\n\n\n\n
Verbatim, Transcribe, Code-mix \u0914\u0930 Transliterate\u0964 Verbatim \u092a\u0942\u0930\u093e record \u0926\u0947\u0924\u093e \u0939\u0948, Transcribe clean text \u0926\u0947\u0924\u093e \u0939\u0948, Code-mix \u0939\u0930 language \u0915\u094b \u0909\u0938\u0915\u0940 script \u092e\u0947\u0902 \u0930\u0916\u0924\u093e \u0939\u0948, \u0914\u0930 Transliterate \u092a\u0942\u0930\u093e transcript English letters \u092e\u0947\u0902 \u0932\u093f\u0916\u0924\u093e \u0939\u0948\u0964<\/p>\n\n\n\n
\u0939\u093e\u0902, \u092f\u0939 Indian telephone audio \u0914\u0930 code-mixed conversations \u0915\u0947 \u0932\u093f\u090f train \u0915\u093f\u092f\u093e \u0917\u092f\u093e \u0939\u0948, \u0914\u0930 native 8 kHz telephone audio \u0915\u094b accept \u0915\u0930\u0924\u093e \u0939\u0948\u0964<\/p>\n\n\n\n
Company \u0915\u0947 Vistaar benchmark evaluation \u0915\u0947 \u092e\u0941\u0924\u093e\u092c\u093f\u0915 \u0907\u0938\u0915\u093e macro WER 16.93 percent \u0939\u0948, \u091c\u092c\u0915\u093f ElevenLabs Scribe v2 \u0915\u093e 19.98 percent \u092c\u0924\u093e\u092f\u093e \u0917\u092f\u093e \u0939\u0948\u0964 \u092f\u0939 company \u0915\u0940 \u0905\u092a\u0928\u0940 evaluation \u0939\u0948\u0964<\/p>\n\n\n\n
Prisma v3.0 \u092f\u0939 \u0926\u093f\u0916\u093e\u0924\u093e \u0939\u0948 \u0915\u093f Indian languages \u0915\u0947 \u0932\u093f\u090f AI tools \u0905\u092c generic English tools \u0915\u0940 copy \u0928\u0939\u0940\u0902 \u0930\u0939\u0947\u0964 Code-mixed \u092c\u093e\u0924\u091a\u0940\u0924, telephone audio \u0914\u0930 Indian numbers \u0915\u094b \u0938\u0939\u0940 \u0924\u094c\u0930 \u092a\u0930 \u0938\u092e\u091d\u0928\u093e \u090f\u0915 \u0905\u0932\u0917 problem \u0939\u0948, \u0914\u0930 Gnani AI \u0928\u0947 \u0907\u0938\u0940 problem \u092a\u0930 focus \u0915\u093f\u092f\u093e \u0939\u0948\u0964 \u0905\u0917\u0930 \u0906\u092a\u0915\u0940 team Indian calls \u0915\u093e transcript \u092c\u0928\u093e\u0924\u0940 \u0939\u0948, \u0924\u094b \u0907\u0938\u0915\u0947 4 modes \u0914\u0930 keyword feature \u0915\u094b \u0906\u091c\u092e\u093e\u0928\u093e worthwhile \u0939\u094b \u0938\u0915\u0924\u093e \u0939\u0948, \u092c\u0938 accuracy \u0915\u0947 \u0926\u093e\u0935\u094b\u0902 \u0915\u094b \u0905\u092a\u0928\u0947 real calls \u092a\u0930 test \u0915\u0930\u0915\u0947 \u0939\u0940 final \u092b\u0948\u0938\u0932\u093e \u0932\u0947\u0928\u093e \u091a\u093e\u0939\u093f\u090f\u0964 India \u092e\u0947\u0902 business mostly phone calls \u092a\u0930 \u091a\u0932\u0924\u093e \u0939\u0948, \u0914\u0930 \u0935\u0939 calls \u0905\u0915\u094d\u0938\u0930 \u090f\u0915 \u0939\u0940 sentence \u092e\u0947\u0902 Hindi, English \u0914\u0930 \u090f\u0915 regional language \u0915\u094b \u092e\u093f\u0932\u093e\u0915\u0930 \u0939\u094b\u0924\u0940 \u0939\u0948\u0902\u0964 \u0910\u0938\u0940 code-mixed \u092c\u093e\u0924\u091a\u0940\u0924 \u0915\u094b \u0938\u0939\u0940 \u0924\u094c\u0930 \u092a\u0930 text \u092e\u0947\u0902 \u092c\u0926\u0932\u0928\u093e \u092c\u0939\u0941\u0924 \u092e\u0941\u0936\u094d\u0915\u093f\u0932 \u0915\u093e\u092e \u0939\u0948\u0964 Bengaluru \u0915\u0940 company Gnani AI \u0928\u0947 28 October 2026 \u0915\u094b \u0907\u0938\u0940 problem \u0915\u0947 \u0932\u093f\u090f […]<\/p>\n","protected":false},"author":1,"featured_media":203,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"_jetpack_newsletter_access":"","_jetpack_dont_email_post_to_subs":false,"_jetpack_newsletter_tier_id":0,"_jetpack_memberships_contains_paywalled_content":false,"_jetpack_hide_featured_image":false,"_jetpack_featured_image_duplicate_dismissed":0,"_jetpack_feature_clip_id":0,"_jetpack_memberships_contains_paid_content":false,"footnotes":"","jetpack_post_was_ever_published":false},"categories":[4],"tags":[20,22,23,24],"class_list":["post-194","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-ai-tools","tag-ai-tools","tag-gnani-ai","tag-prisma","tag-speech-to-text"],"jetpack_sharing_enabled":true,"jetpack_featured_media_url":"https:\/\/akshaymaurya.in\/blog\/wp-content\/uploads\/2026\/10\/thumb-prisma.jpg","_links":{"self":[{"href":"https:\/\/akshaymaurya.in\/blog\/wp-json\/wp\/v2\/posts\/194","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/akshaymaurya.in\/blog\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/akshaymaurya.in\/blog\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/akshaymaurya.in\/blog\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/akshaymaurya.in\/blog\/wp-json\/wp\/v2\/comments?post=194"}],"version-history":[{"count":1,"href":"https:\/\/akshaymaurya.in\/blog\/wp-json\/wp\/v2\/posts\/194\/revisions"}],"predecessor-version":[{"id":195,"href":"https:\/\/akshaymaurya.in\/blog\/wp-json\/wp\/v2\/posts\/194\/revisions\/195"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/akshaymaurya.in\/blog\/wp-json\/wp\/v2\/media\/203"}],"wp:attachment":[{"href":"https:\/\/akshaymaurya.in\/blog\/wp-json\/wp\/v2\/media?parent=194"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/akshaymaurya.in\/blog\/wp-json\/wp\/v2\/categories?post=194"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/akshaymaurya.in\/blog\/wp-json\/wp\/v2\/tags?post=194"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}
External source: CIOL \u0915\u0940 report<\/a><\/p>\n","protected":false},"excerpt":{"rendered":"