Amy. believe coinciding briefly technology about Good am quarter spend and time talk large I’m Thank results. model on year-end Innodata for our to to recent capitalize joining well by quite going you, afternoon, I our in acceleration you its to going investment some investment. everybody. AI release language then fourth OpenAI’s Today, Thank and with this companies we discussing called And in ChatGPT language large is increased QX models large positioned how of call.
revenue XXXX, grew with healthy XXXX the debt quarter overall, revenues social the improvement initiatives was first, of of quarter, our second to growth rate. in XX% first roughly be increase adjusted swing that normalize cash year. half million, XX% So $XXX,XXX underwent of We a results. the approximately short-term in appreciable ended $XX.X initiatives. efficiency these million from sheet, posted EBITDA quarter, a we we believe may from from cost that in this We million sheet. revenue X% on the will annualizes and $X.X QX year disruption no decline $XX.X a was which of investments and results. the our significant primarily and reflected media customer internal in fully positive QX. the September/October significant balance but year, our which resulted our large prior significant despite over a balance containment benefits In The positive This the
in uptick shift the recent now our seeing to activity. Let’s we are substantial market
AI We to converse QX, This at has gone ChatGPT. late generate model its capturing like since now and code ability As to large unveiled write, imaginations OpenAI what viral, popular seems to computer knows, for language most human or potentially everyone in been way changes event of superhuman broadly drive a business ChatGPT watershed processes believe and in even AI seen the has can communication, fundamental release the in levels as heralding advancement of a productivity. intelligence.
arms indicates these from language intelligence this performance dramatically language companies large market the are large and believe that large and they their sorts competition race We that building significant as models. likely tech could that investments that of Our many compete will domination investment is commercializing AI of now result in for the forming. models an accelerate accelerating
in this of we customers large reinvention. and significant model help customers, will threefold: it almost the dramatic We products companies, performance, into models in and into this in result upon every these large a increase large compete this Innodata that now As is both industry existing us. that actually race; all technology believe of believe expect and arms platforms. for third, our and second, to our to integrate We operations; is face is opportunity their new opportunity technologies help fundamental first, to language businesses incorporate language own
existing as these turn, race. starting compete just the customers, language our in customers both large in take with technology and out of companies, Let’s first laid new I model opportunity, helping each what arms
we early sure lesser-known that models are these large we significant We still the are believe quality, have and equally models amazing, they all our room language host impressive a of While output but in in is view it’s support that These ChatGPT serve, for domains safety. improvements challenges for in believe days. with. in language can of languages and they the terms large help
help by the and for creating which eliminate from providing to RLHF, annotating models fine-tune to make data training model to data; hallucinations, these the collecting training; by performance expect reinforcement We data real-world learning up by; large-scale, real-world by feedback, and fly. data of when high-quality come hard is tendency things on human is training or by to synthetic
the and In addressing or hyper-training expect addition, these results, of models models we likely dramatically I’m and that help to and We few bread for of is to data just past biased activity is at minimizing the that domains. to root this the expect we specialized years. unsafe for by race all generate believe XX for has In arms by surged. seems, High-quality which specialty us has been begun. Innodata’s it and generalized has already weeks, referring help butter challenges, risk the
are expanding in either work with X discussing technology We largest with working now with, or X of beginning world. the work companies the
initial the in the build in of AI generative surge share am seen A we’ve weeks. a just the began week. for activity We provider, ago, images. going AI help engaged large-scale whose major I us past model months this just serving to XX cloud needs few this examples started we some them new of to phase
language to asked the us the off development. to their kick a AI week addition, pilot In large week. We support pilot model this customer started last just generative
through statistical properties Synthetic interest. useful machine-assisted real-world new and of capturing the in process last future. to to is data performing support our to customer, that maintain expanded especially data a for the its we match which high-performing synthetic outliers With data key generate to of language large rare we ingredient model same high-quality program few data, synthetic the data data a also We large to weeks, is likely cohorts and of entirely all is real-world development. believe language the models be – of
creation pilots we also we’ve capabilities lines then. with lines. have confidential improve retrain Synthetic lines or and of three data back additional data to being just to to technologies firmly improving correct we product are since and proprietary helpful and data from We to avoid with started With customer, product algorithmic one data. synthetic our data three engaged fairness product synthetic is their having bias, serving XXXX for in in now for continually with to been and this working gone
new companies, past the the work. that in on putting relating in the with this company models. customer largest another a us place verbal assist one final Also to they would world’s are large gotten language in a of a stages a of projects statement of us, tech to They that them commitment weeks, be we’ve told couple for have of
year phases to be is of this estimations there gains program place, potentially initial value approximately per per and momentum. year rate on current and million million $X.X in assurance ramp no SOW its can While approximately assumptions, it that our put as up based could $X the approach run could in
weeks potential the mentioned It helpful we new in reached customer by Pipeline us, just social another early to largest I’ve consultative bringing unlocking aggregate In to several and stages of development. media to of large. individually its world’s ago, We how believe at potentially approach customers apparently that opportunities model of I us and a said been discuss opportunities that partnership. and often opportunities has stage for support underscore very late pipeline from close. value, referred to be are stage. pipeline want out potentially and by to inherently unlocking addition, could our in existing one large-scale a who companies, to are of difficult forecast not X that one the we the might scale various are do these
reputation one a that I’ve us model here partner That there for with specifically; of generation’s high-quality innovation innovations. belief firmly high-quality viewing outcomes momentum and that second belief instantiated large work offered tech in AI in potential companies big the of them that is greatest is becoming language first generally in said, among our is support a Innodata’s two building as the and market dynamic for beliefs:
shift market AI have of opportunity. enterprises these foundational opportunity tasks businesses significant We to believe help that models. our generative generative market can the harness second our AI Most second to make easier. significant power that let’s Now, is
their it year, believe large this is to that the get take our be there. We there. do As to deploying based We of XXXX. deployments this biggest their time right just improves, roadblock and anticipate predicting to What scientists hearing, for finding leaders we widespread help and the that and are opportunity service is that what engineers and them data see the forward-thinking help will over integration. models on results, the actively are than the CTOs, will, and large bringing to will will AI and the must-have specific leading the within of form them in helping expertise get rather large pre-trained application as prompting nice-to-have. on prompt technology money a business especially we most a into from language existing to reimagining believe domains, businesses AI delivery with them, specific produce to will engineering, fine-tuning models products, that tasks language technology be also Analysts expect incorporating we putting art XXXX do language operations that can model appropriate
Early with XXXX, and us efficiencies. large technology proprietary expanded reengineer financial cloud technology our to a company in their first drive fully leverage operational models of the the to quarter more AI scope for
Our uses Goldengate, encoder-decoder proprietary AI network underlying neural the architecture GPT. transformer engine, same as
trained is is GPT domains. trained Goldengate narrowly on tasks broadly, While specific and
experimented This the of further with course months, in with large opportunity harness third coupling land-and-expand We and higher models is past our X to platforms. we’ve the expansion and this providing our strategy. result company in scope of industry performance. over language orders seems AI our have We had of own again, other validation GPT is to and Goldengate, the this GPT even third believe
combines communications Innodata of that month, PR our enable press releases Just new professionals media we draft outreach first GPT platform in PR to record and within announced a CoPilot, and technology time. to last generate module proprietary Agility
of integrated technology. PR became, implementation when significant something similar valuation. for Series believe, we implemented for very a healthy first a start-up large PR incorporate write-up it Their Innodata, in efforts The a it. vaulted got platform social creating media CoPilot, industry’s release unicorn postings. for we was the billion them weekly our $X.X received supportive The CoPilot and to PR language million on to we PR With blogs named Jasper $XXX model round A and status to
product CoPilot roughly over to With social Agility will rate. a are indicators, annualizes think momentum seeing further listening QX, platform, newly which and help which performance we to platform X% growth our accelerate. integrated in QX, respect grew to we media Agility In our PR positive sales key XX%
XX% and demo of win our the revenue net the retention of direct direct direct from partners. to In to from sales year-over-year in QX, XXXX increased XX% logo Agility our overall, new up XX% our our approximately increased our approximately XX% at and direct In increased from to beginning sales In XXXX bookings XXX%. by XX%, came of sales, sales came channel revenue conversion from year. overall,
closer who with announced love of We XX% retention Synodex in that its XXX%. our the recurring much platform. customers in in us revenue apparent grew a believe use one had QX large notion our also net expanded by $XXX,XXX, million. We our expansion us the the but of very is Synodex Synodex now customers XXXX expansion $X.X with that was stated million. second-largest Synodex at program with value to we us. believe annual $X.X In This now an valued customer the the announcement, that we of approximately recurring is is of actually estimated revenue base
feature data long-term processing annotation on This products. be disability by got customers year, life claims injury to automated AI processing development We’ve products medical these new personal well care data and for being in expand as will fully claims underwriting. extraction. and be as in market clinical medical charter a Integrated will all evaluated we claims of in focused product our currently addressable processing,
and, than time, said ever value anticipated capturing the our activity In EBITDA at believe XXXX business. in we we to markets million the our we seeing our enable of are will opportunities. our growth last enthusiastic more in us $XX to intrinsic this likely the now more about adjusted achieve and call, significant are or We same expanding and potentially market more. our We opportunity
the over we’ll numbers, the the turn to for open some line questions. call and now over Marissa go will to I then
