@JohnnyM @jcmleng

I think that AIM's iCAP network and having trusted embeded hardware/software encoders that enable automated coordination of workflows are a moat for their current market ie live captioning.

But @JohnnyM raises an important risk for future growth of LEXI into other areas such as LEXI VOICE and LEXI Audio Description which will be important if AIM are to reach their 5 year growth target of $60m of EBITDA in FY29.

@Shapeshifter , thought more deeply on your comments as it bothered me somewhat and I wanted to refresh my understanding of AIM's approach in this space. I went back to my notes on the SM Meeting and the AIM AGM - extracts of the notes plus a bit of context around the key statements (bold, italics) are per below.

It does appear to me that:

• the approach is to extend the current moat for live captioning into Audio Description and Voice - that makes complete sense in terms of leveraging your moat into adjacents.
• AIM does a lot more to Audio Description than the translation bit - the end to end process is an involved one which per AIM, involves ~25 hours of human effort, which LEXI Audio Description will fully replace and improve on - the capability to contextualise is key
• the market for Audio Description is still the big-end media/government customers
• AIM does not compete from a technology perspective but integrates whats out there - if VLC further expands capability, it could well be used in the future AIM toolkits/ecosystems

Will keep the VLC risk on my radar. Never say never for sure, at this stage, but I don't quite see the risk impact.

--------------

SM Meeting

The big prize if the $69b grey space which is to use AI to deliver products that involve voice and dubbing - 2 products to be delivered in Oct

• Audio Description - equivalent captioning-for-deaf people, for blind people - describes the visual element of the storytelling that a blind person would not pick up but are really important to the storyline and which is not actually verbally said eg. Clark Kent goes into the phone booth, spins around and then comes out with ....
• Traditional process for Audio Description is 25 human hours for 1 hour of program content
• Legislated minimums for audio description in the UK and Europe, not yet in Australia but legislation is about to be introduced, currently about 14 hours a week on Govt-funded channels eg. ABC, SBS
• Media industry has no appetite to add extra costs for struggling media organisations - frustration in the blind community on how to get access to these services
• AIM’s offering, LEXI Audio Description, is fully automated - uses Gen AI not just to read audio, but to read the vision, understand the text, and to make the decisions that previously was made by humans ie. Identify what were the salient elements of the visual story telling that were not described in the text, find a gap in the audio to insert the description into, write the script for that, have a voiceover artist read it in, mix the audio, and have it available as an audio track - this is what makes up the 25 hours today
• Expect LEXI Audio Description to massively increase the TAM because it will be a case of “why wouldn't I just do this”? as the economics become very compelling
• While the blind community is not a large one (deaf/hard on hearing population is about 10x the blind population), the market is actually the media organisations who is or will be compelled to provide audio description content

Competition

• AIM does not compete from a technology perspective - it integrates and USES the evolving technology to improve its offering within their customer workflows
• AIM DOES NOT DO THE AI ENGINE - it does not invest in speech detects engines
• Invest instead in the API calls that allows AIM to connect their customers signals with the latest and greatest AI engines - not betting on any AI text engine to win, have and use dozens of AI text engines in the Cloud

AGM

The addressable market for the LEXI VOICE is significantly larger than for LEXI captions. By continuously incorporating the latest advancements in AI, we continue to offer market leading solutions, powered by the same architecture that has delivered success in the US Broadcast market, further strengthening and extending our defensive moat anchored in LEXI over iCap with secure access to customer data.

A major milestone was reached in November 2023 when LEXI Live overtook the quality of legacy human captioning achieving an average score of 98.7% on the NER scale, which is an accepted industry method for determining accuracy of captions. In the last 12 months, the quality of LEXI Live has continued to increase, with results now in excess of 99.1%. Because of this sustained quality improvement momentum in LEXI, less than 20% of our business will involve human curation by December 2025.

Over the period to FY29 we anticipate LEXI superseding humans, as it first did with LEXI Live Captioning in 1H24) in all other AI-driven language products serviced by our LEXI Toolkit – including the category that dominates the language services market by value - VOICE.

@jcmleng yes, that’s how I see it. There are no shortage of offerings that can plug a mixed audiovisual track into software connected to an LLM tuned to natural language in any language. That’s a zero moat.

But as I understand it, $AIM is about integrating their modules and software into the broadcasters work flow within the studio where the multiple feeds are controlled and mixed. And that’s more complex.

By all means, we all need to keep a track on these developments. But AI driven captioning of a mixed output doesn’t present a threat. Or at least a new threat.

Good notes @jcmleng and that does align with my understanding of AIM from the SM meeting and the CEO address after the AGM.

AI media have an incumbency advantage. Their trusted live broadcasting captioning solution is deeply woven into the workflow of many US broadcasters. AIM's LEXI ecosystem is a great lever for it's new products however obvioulsy they will face competition.

The VLC media solution that was posted by @JohnnyM sounds similar to AIM's  LEXI Recorded that AIM launched in FY24. From that article, "VLC Chief Marketing Officer Natacha Holtzhausser says the AI-generated captions and subtitles will open up accessibility, no matter the language, for downloaded movies and TV shows, lectures, and live streams." This will be a free solution and the VLC Media Player has had over 6 billion downloads since inception.

So although not currently a direct competitor with AIM, VLC's solution does demonstrate that the barriers to entry are not that high and it indicates to me where risk to the AIM growth thesis may come from. It may not be from AIM's traditional competitors.

However AIM has and is continuing to develop an impressive ecosystem. This will not easily be overcome.