MC still, ML markup_PitchP DRAFT4.png

Movie still provided by MovieComm.

Mezzanine.ai brings together single-point integration and automated orchestration of readily available pre-trained machine learning models from AI industry leaders including Google Cloud, IBM Watson, AWS and Microsoft Azure.


In The News

EQD08 - JUN19 - CoverStoryv2 _ COVER PG.

10 Most Innovative Companies in ML:

"Channel Chief Works to Democratize Machine Learning for Businesses by Focusing on Building the Bridges

to ML Models"

SiliconRev JAN20 - CoverStory _ COVER PG

10 Best AI Companies to Watch 2020:

"On an Ambitious Mission to Bring the Incredible Benefits of Artificial Intelligence to Every Business: mezzanine.ai" Jan 2020


Mezzanine.ai is a new approach for businesses looking to engage with machine learning. It empowers your data to be seamlessly and securely evaluated through multiple pre-trained machine learning models through a sophisticated, wizard-driven orchestration and automation platform.

Mezzanine.ai delivers on the promise of bringing AI and machine learning capabilities directly to businesses, with absolutely no need for in-house AI or machine learning knowledge, specialized technology or data scientists. By leveraging a multiplicity of machine learning models simultaneously, the detail and volume of data insights increase exponentially.



Through its comprehensive application integration library, mezzanine.ai’s underlying patented platform establishes secure synchronizations between business applications and data, whether hosted or on-premises, allowing companies to instantly leverage the power of AI and machine learning.



The mezzanine.ai ML marketplace consists of hundreds of available application integrations across a wide range of hosted and on-premises solutions, as well as a large catalogue of curated ML models. A wizard-driven interface enables automated data flow from your application to the machine learning model of your choice.



Through sophisticated orchestration and automation capabilities, mezzanine.ai directly facilitates the transmission of data between multiple ML models and offers follow-up orchestrated actions such as redactions, archiving, or further analysis.


MEET mezzanine.ai


Mezzanine.ai was developed as a stand-alone solution through the utilization of Geminare’s award winning and multi-patented orchestration and automation platform. As a leader in delivering data curation, resiliency and IT orchestration solutions, Geminare has helped advance global service providers with market leading orchestrated Cloud solutions, an example of which is the award-winning Resiliency Management Platform (RMP). The RMP powers Magic Quadrant Leaders among many of the world’s leading Service Providers including NTT Communications, Cable & Wireless, Liberty Global, Recovery Point Systems, TierPoint LLC, Iron Mountain, and many others. With mezzanine.ai, Geminare is leveraging its capabilities to power the next generation of machine learning solutions for business.


Mezzanine.ai is democratizing machine learning for businesses by enabling them to access whatever models are best suited for their needs from whichever platform they chose to use, whether it’s just a single pre-trained model or all of them at the same time. With mezzanine.ai, businesses can test and evaluate the results of their data as they are run through multiple pre-trained ML models from a single interface, almost instantaneously. The wizard-driven interfaces within mezzanine.ai accelerate access to readily available pre-trained machine learning models from AI industry leaders including Google Cloud, IBM Watson, AWS and Microsoft Azure, enabling immediate insight into enterprise data and facilitating smarter curated business outcomes.


Mezzanine.ai is a SaaS hosted platform incorporating a comprehensive custom developed application integration library. These integrations provide the foundation for secure synchronizations between business applications and their associated data, both hosted or on-premises as well as a curated library of readily available pre-trained machine learning models, allowing companies to instantly leverage the power of machine learning.

AI Gl,Az,IBM,AWS (no border).png


Industry: Media & Entertainment

MovieComm helps thought leaders inspire and engage others by using Hollywood movie clips to make communications come alive. Customers work with MovieComm to obtain clips with specific references or themes for projects ranging from embedding high-value content into multiple communications platforms including internal communications platforms, LMS systems, PowerPoint presentations, email and text messages, to using iconic film clips to enrich communications. With a published goal of over 1,000,000 documented clips to provide to the market, leveraging the power of AI and machine learning was a clear necessity and represented a potentially big win for MovieComm.

MC still, ML markup_Lebowski DRAFT.png

“All machine learning models are not created equal. MovieComm is now able to deliver a vastly superior end-user experience, and fully leverage the value of our films’ assets. Overnight, we’ve gone from a usable database of tens of thousands of searchable items to millions, and from a cataloging time of weeks to just minutes. Our experience with mezzanine.ai capabilities has been invaluable.

Scott DiGiammarino, CEO MovieComm

Movie still provided by MovieComm.

MC still, ML markup_Apollo13 DRAFT2.png

Movie still provided by MovieComm.

MC still, ML markup_NottingHill DRAFT.pn
MC still, ML markup_NottingHill DRAFT.pn

Movie still provided by MovieComm.

MC still, ML markup_PitchP DRAFT4.png

With mezzanine.ai, multiple ML models are leveraged in parallel, providing vastly superior and a larger quantity of results. 

Without mezzanine.ai, each data set must be curated and formatted per ML model, and sent to each model one at a time.

Results Figure 1: A MovieComm sample clip run through multiple ML models, with results orchestrated through mezzanine.ai

U-FOD-1 Portal - ML results Rekognition.
Google NatLang_SIMPLEtest results1 edits

Results Figure 2: The same MovieComm sample clip run through a single provider’s hosted ML models

Movie still provided by MovieComm.

Use Case - 


Technical White Paper - 

Learn more about how mezzanine.ai and machine learning worked together.

Industry: Customer Support & Sales

It’s all about reaching your customers, listening to them, and understanding what they are telling you. But how do you interpret and analyze the endless hours of recorded phone calls or growing repository of online chats? How do you know which products are being discussed, whether a competitor’s brand is mentioned, and if customers are happy with a recent promotion?

The key to using ML models is extracting all of this information and sentiment from your existing data and eliminating the need for manual intervention. From there you can ask business questions of the data to follow trends, identify problems and common concerns, and make educated predictions for future campaigns and strategies.



Speaker ID 1: Hello, how can I help you?

Speaker ID 2: I’m looking for help on ProductA. I’m not happy with it so far. And how do your promotional products compare with TheOtherGuys?


Speaker ID 1: Certainly, we can assist you. And we just released ProductB at a discount, so we’re less than our competitors.


Speaker ID 2: Great, I’m happy to hear that, tell me more.


Identify if the overall sentiment of a conversation is positive / neutral / negative, and break down individual sentences to analyse the sentiment trend over the course of the interaction.


• The overall sentiment was neutral.

• Further analysis of the interaction revealed that the interaction started off with a mild negative sentiment, while ProductA was discussed.


• Sentiment improved after ProductB was discussed and the interaction ended on a strong positive tone.


• Overall trends from multiple calls show that sentiment is worst on Mondays and best on Fridays.



Identify if specific keywords are included in order to track product placement and review sales strategies.


• ProductA was discussed at the beginning during the negative interaction.

• ProductB was discussed at a later point after which the sentiment was positive.

• ProductC was not mentioned.

• The customer did not mention cancelling any services.

• 1 competitor was mentioned by the customer.


Share your call center transcripts around the world.

Parleur No. 1: Bonjour, comment puis-je vous aider?


Parleur No. 2: Je cherche des informations sur ProductA. Je ne suis pas content avec ça jusqu'à présent. Et comment vos produits promotionnels se comparent-ils à TheOtherGuys?


Parleur No. 1: Absolument, nous venons de publier ProductB à un meilleur prix, nous sommes donc moins que nos concurrents.


Parleur No. 2: Excellent, je suis heureux de l’entendre, dites m'en plus.

ナレーター1: こんにちは、どのように私はあなたを助けることができますか?


ナレーター2: ProductA に関する情報を探しています。 私は今のところ満足していません。 そして、あなたの販売促進製品は TheOtherGuys とどのように比較されますか?


ナレーター1: 確かに、私たちは ProductB をより良い価格でリリースしたばかりなので、競合他社よりも少ないです。


ナレーター2: 素晴らしい、それを聞いてうれしいです、もっと教えてください。

Industry: Security & Compliance

Identifying, protecting and redacting sensitive data is more than just an important feature. For most industries, it is a requirement both legally and to protect critical customer data. With data repositories scattered across call centers, email exchanges, support cases, internal and external clouds, and file repositories the ability to identify and protect this data has become a monumental task in virtually all industries.


Machine learning is uniquely capable of identifying sensitive data elements like credit card numbers, names, social security numbers, identifier numbers, phone numbers, and credentials. ​Using mezzanine.ai to orchestrate the scanning of your data, wherever it might be, in order to find and recognize certain data types, and then act on the information to redact, delete or encrypt private data, provides automated compliance for virtually all industries.

sample customer data.png



Mezzanine.ai has uniquely integrated capabilities enabling the association of similar data characteristics across multiple ML models. This orchestration capability empowers insights into data that are not possible through the use of single data analytic results.


Identifying concerning data characteristics in minutes versus days allows you to realign searches and narrow your focus virtually on the fly. With mezzanine.ai you can immediately see which data needs particular attention and from which data source. 



The power of mezzanine.ai’s orchestration engine becomes immediately evident as results are curated and data is protected through redactions, archives, removals and virtually any other form of orchestrated request – all automated and sequenced.

data relationships.png
report results DLP.png


If you’re new to Machine Learning and don’t yet understand what benefits might accrue, that's ok, we can help.  Our team has deep knowledge and experience in this industry and can offer unique insights into how Machine Learning can help your business.

So, if you’re interested in learning more and/or would like a demo of mezzanine.ai, please contact us and we'd be happy to speak with you.

Tel: 1-650-319-8577 x 1004
4962 El Camino Real, Suite 103
Los Altos, CA 94022 USA


To schedule a product demo with one of our product consultants, please fill in your contact details.

Success! Message received.



Joshua Geist Dec 2013 B&W.png

Joshua Geist,


  • Grey LinkedIn Icon

Joshua Geist, CEO of Geminare and mezzanine.ai, a career entrepreneur, presenter, multi-patent holder, technology industry commentator and expert is a recognized innovator within the IT Orchestration Market.


Geist is a recipient of numerous industry accolades including multiple CRN Channel Chief awards, Gartner Cool Vendor designations, Most Innovative Solution winner from CIR international community, and lead contributor in the positioning of multiple Gartner Magic Quadrant Leaders including global telco giants such as NTT Communications and Cable & Wireless. Geist has had direct responsibility for spearheading the launch of some of the largest and most successful Resiliency offerings in the market.


Leveraging the core technology strength of the Geminare platform as well as the breadth of experience launching Cloud Solutions to the enterprise, Geist has been pivotal in launching mezzanine.ai, another industry first SaaS platform which enables enterprise organizations to de-risk their AI initiatives by leveraging existing machine learning models in a simple, orchestrated and secure fashion.


Geist previously founded and was CEO of MindBytes (acquired by Munimentum Group), a professional services organization focused on delivering technology best practices to small and medium-sized organizations. There he acted as the outsourced Chief Technology Officer and advisor for over 200 small and mid-sized organizations over the course of 12 years. Prior to founding MindBytes, Geist had the privilege of working for Apple Computer for many years as their Technical Training Manager where he developed and delivered network, system, sales and operational training programs across Canada.


Geist holds a degree in Physics, sits on multiple government and technology industry boards, and is based out of Geminare's Silicon Valley offices.

Mike Larkin Jun 2012 B&W cropped.png

Mike Larkin,

Corporate Advisor

  • Grey LinkedIn Icon

Mike Larkin is an experienced executive whose career spans over 30 years, primarily in the telecommunications and emerging technology fields. Mike’s senior leadership roles have included tenure with Bell Canada as VP Sales and Marketing as well as various start-up ventures including iNet America, BCE Emergis and WorldLinx – an early-stage Internet company that grew from zero to $50m in revenue under Mike’s direction.

Mike is the co-author of two best-selling business books, entitled ICE and CARE. He is married with twin children, and resides in Oakville, Ontario, near Toronto.

Susan Bantin d850-8505297 bw.png

Susan Bantin,

Dir Bus Operations

  • Grey LinkedIn Icon

Susan Bantin brings over 25 years of operations, program and project management experience from a variety of industries including Cloud Computing, Electrical & Energy Engineering, and Healthcare. Susan leads Business Operations in managing the mezzanine.ai Artificial Intelligence team as well as ongoing research and development projects, and strategic product lifecycles. Past experience includes engineering and professional services roles at multiple start-ups and enterprise technology firms. Most recently, Susan managed the multi-million dollar Engineering Operations department at a Toronto-based green-tech power company, supporting customers across 30 countries, and successfully doubling the team over three years. Additional career paths include paramedicine with Toronto Paramedic Services and Lab Instructor at Centennial College’s Paramedic program.


Susan is a P.Eng. and holds a degree in Engineering Physics. Her interests include AI in Healthcare, jazz, and mountain biking, and in her spare time, she is learning to play guitar.