What makes a good AI product? What elements do we look for to predict commercial success?

Human brains are amazing things, capable of things that computers just can’t do. AI is amazing too. Combining AI with human smarts elevates human ability to new levels. Our work focuses on using artificial intelligence to augment human capabilities and make workforces more productive.

At Red Marble, one of our offerings is to help clients develop and commercialise AI-powered products. We also develop and license our own products, always with an overarching focus on how AI can enhance human performance and workforce productivity.

Embedding AI into a product allows us to solve one specific business problem extremely well. And as the software learns, the solution improves over time.

So what makes a good AI product? What elements do we look for to predict commercial success?

How we define an AI-enabled product

  • A set of code and algorithms solving a specific, clearly defined, repeatable business problem in an area of business value.
  • Has defined inputs and outputs and can be deployed as a service
  • Includes algorithms and machine learning with feedback methods to learn, and becomes more intelligent the more data it consumes.
  • Based on unique and specific data which is usually not freely available (and more data creates a barrier to entry).
  • Can be deployed for sufficient time for the model to learn and to fulfil its potential.
  • Requires minimal (< 20%) configuration for a specific customer and doesn’t require large consulting  effort to implement.
  • Harnesses the power of technology to augment human capabilities

That last point is fundamental to what we do. We’re here to enhance human performance

and elevate human productivity - not replace people with machines.

It’s important to note here that a lot of the work in developing AI-based solutions is not machine learning; around 75% of the work relates to software engineering, data cleansing, data engineering and similar tasks. The actual model development, although often the most valuable part, is relatively small (Bastian Huang from Osara in the US describes the issue nicely here.). So it’s crucial that we’re factoring in those other tasks when we think about creating a new product.

There’s a couple of key considerations which are important to contemplate up front.

Start with the business challenge, not the technology

There are two ‘non negotiables’ before we develop a product:

  1. It needs to focus on a clearly-identified and specific business problem
  2. We need to be able to generalise the model across multiple customers

If those aren’t in place, the work should probably be considered as a software development exercise with intelligent algorithms part of the solution, but not a product offering.

We like a product to have a single ‘job to be done’, which keeps us focused on the business challenge, and how we solve it in a generalised way.

Work out how you will tackle the ‘long tail’

A big challenge of developing any product incorporating AI and machine learning is the ‘long tail’ - the large number of items which exist in small quantities, rather than the smaller number of popular items.

The concept is described nicely in this article:

“Supervised learning models tend to perform well on common inputs (i.e. the head of the distribution) but struggle where examples are sparse (the tail). Since the tail often makes up the majority of all inputs, ML developers end up in a loop – seemingly infinite, at times – collecting new data and retraining to account for edge cases.”

Many machine learning models can have a disproportionately long tail and the cost of training the model can rise exponentially, often making the project unviable without deep pockets. That’s why we prefer models where the ML does the heavy lifting, managing the bulk of cases, and allowing human team members to manage edge cases. That way we can still provide a great amount of value while keeping costs lower.

Our product journey

We’re currently working on a number of products, where we focus on incorporating intelligence into software, but always with the human at the centre. We view every opportunity through a human-centric lens. Being so attached to people may sound odd for an AI company, but it helps us (and you) reach our core goal: liberating humans to achieve more.

Want to know more about our AI-enabled product development approach and see if and how it can help you? Get in touch and let’s discuss your biggest challenges.


How AI can augment prediction and human decision making

We spend much of our time here at Red Marble exploring ways that AI and machine learning can elevate human performance. One area of particular interest is how AI can augment prediction and human decision making.

But before we can design AI, we need to understand the differences between how software and human brains make decisions.

How do humans make decisions?

Humans make thousands of decisions every day; subconsciously combining inputs from multiple parts of the brain, combining real-time data with historical information from our memory and blending rational thought with emotional cues.

In “Thinking, Fast and Slow,” Kahneman and Tversky explain how fast, instinctive and emotional decisions are blended with others that are made logically and deliberately - all with an assessment of risk, probability and judgement based on experience.

This is a pretty sound way to make decisions - however, it takes time and the quality of the decisions can depend on outside factors (ie the person’s health or mental state).

How does machine learning make decisions?

Machine learning (ML) models aim to emulate elements of this decision making, but clearly some areas are more accessible than others.

ML reflects the ability for software to ‘fit’ a particular model to a set of data by applying specific weightings to different parts of the data (called ‘features’). It uses this model and applies the weightings to extrapolate that data in order to predict future events.

A machine can process vast amounts of historical data to make its predictions, with a defined probability, based on past data - but it can’t apply an emotional lens (yet) to those decisions, and it struggles where context changes.

Where machine learning works well

A nice application of ML is to model the rational human decision making process and to make those decisions at scale. Let me share an example.

We recently worked with a client who was making predictions about stock levels of spare parts for machinery. Were they holding enough in stock? When spare parts were ordered, would they arrive on time? They needed to know they would have the parts when required.

Looking at the “health” of each material made the assessment fairly intuitive and simple for the human. A track record of late deliveries from suppliers, highly variable stock levels in the warehouse, parts being used for breakdown (rather than planned maintenance) all lead to a fairly simple judgement by the human worker.

The challenge is applying that process across 800,000 materials every day. Clearly not something a human can do.

Applying an ML model here does the heavy lifting superbly and creates a list of priorities that the human can work through and apply judgement to.

We apply similar models in our AI-driven employee engagement and digital adoption work. The software can analyse many users and predict which information is most valuable to help each individual use their technology to its full potential and to succeed in their role. It models the human analysis, and applies it at scale.

Where do ML models fall short? 

Machine learning takes historical data but can lack immediate context. Here’s another example.

We help one of our clients predict customer conversion for certain products and prices, helping them maximise margin. The model is trained on historical data, but within the context of COVID, historical data is not reflective of current buying behaviour. Companies need to adapt their models to make use of real-time context and the most recent data as it's coming in.

Machine learning can also fail to understand the behavioural aspects of decision making. Human decision making has lots of nuances; emotion, the time available, the paradox of choice as outlined by Barry Schwartz to name a few. AI models can be trained on aspects of these - and they will get better at this as technology improves - but for now will struggle to put it all together into cohesive reasoning.

When is emotional decision making a problem? 

There are areas where prediction models can expose harsh realities which the human mind might wish to overlook.

One of our areas of work is predicting the success of corporate projects. By tracking the data throughout the project and applying a model trained on historical information, there are clear ways of predicting which projects will succeed, and which will fail to deliver the predicted business benefits. However, many projects are “pet projects” of a particular group or person; and the human mind may already have an ideal end state in mind which will bias the decision making.

An ‘emotionless’ ML model will have no issues exposing the hard facts: “This project only has a 30% chance of delivering the desired benefits” quickly cuts to the reality and can save companies countless time and expense.

Human decision making + AI: what’s the best way to combine them?

Humans are amazing, and the ability to make complex decisions pretty quickly is significantly ahead of our software abilities. However - add in the capability of AI to apply models rapidly, constantly and at huge scale to the capability of human decision making, and you have some very exciting possibilities.


What IS GPT-3? And why do you need to know?

At Red Marble, our core belief is that artificial intelligence will transform human performance.

And as part of our everyday work, we’re continually coming across (and creating) ways that AI is improving workforce productivity.

Our projects generally fall under five technical patterns of AI: prediction, recognition, hyper-personalisation, outlier detection and the one I’ll talk about here...

Conversation and Language AI

Broadly, conversation and language AI deals with language and speech. There are 3 aspects to this pattern:

  1. The ability to have a conversation with software, either via text or voice. Common examples of this include Alexa or Siri, but we’re seeing an increasing number of voice-based interfaces within enterprises.
  2. The ability to understand language and analyse it; for example, we recently worked on a project where we analyse text in work notes  to understand if any contractual clauses may have been triggered.
  3. The ability to generate language - to create a natural language narrative based on input data, for example auto-generating a project status update narrative based on data collected.

There’s been a huge advance recently in natural language generation. It’s based on software called GPT-3 (Generative Pre-Trained Transformer, version 3) developed by California-based AI research centre OpenAI.

This technology was flagged in a research paper in May, and released for a private beta trial in July 2020.

What IS GPT-3?

GPT-3 is a ‘language model’, which means that it is a sophisticated text predictor.

A human ‘primes’ the model by giving it a chunk of text, and GPT-3 predicts the statistically most appropriate next piece of text. It then uses its output as the next round of input, and continues building upon itself, generating more text.

It’s special primarily because of its size. It’s the largest language model ever created, trained using around 175 billion variables (known as  ‘parameters’ in this context). Essentially it’s been fed most of the internet to learn what text goes where in response to certain input primes.

What can GPT-3 do? 

Some beta-testers have marvelled at what it can do - medical diagnoses, generating software code, creating excel functions on the fly, writing university essays and generating CVs just to name a few.

Others have rejoiced in posting examples showing that GPT-3, though sophisticated, remains easy to fool. After all, it has no common sense!

https://twitter.com/raphamilliere/status/1287047986233708546

https://twitter.com/an_open_mind/status/1284487376312709120

https://twitter.com/sama/status/1284922296348454913

Why is GPT-3 important? 

For now this is an interesting technical experiment. The language model cannot be fine-tuned - yet. But it’s only a matter of time before industry-specific variants emerge, trained to skilfully generate excellent quality text in a specific domain.

Any industry where text based outputs or reports are generated - market research, web development, copywriting, medical diagnoses, property valuation, higher education to name a few - could be impacted.

Is GPT-3 intelligent? 

This is the big question for us and cuts to the essence of what we think makes for great AI.

In our view, GPT-3 is great at mathematically modelling and predicting what words a human would expect to see next. But it has no internal representation of what those words actually mean. It lacks the ability to reason within its writing; it lacks “common sense”.

But it’s a great predictor of what a human might deem to be acceptable language on a particular topic, and - we believe - that means that through all the hype, it’s a legitimate and credible model. We’ll be keeping a keen eye on it!


Hyper-personalisation: the secret to “one-to-one” communication

At Red Marble, our core belief is that AI will transform human performance. Within an enterprise, how can AI transform workforce productivity?

Most of our work falls into 5 technical patterns: prediction, recognition, conversation and language, outlier detection or the one I’ll talk about here…

Hyper-personalisation.

This is the ability to treat every customer or every employee as a unique individual. It enables you to move to “one-to-one” marketing or far more focussed employee engagement.

You’ll probably have experienced the recommendation engines in Amazon or Netflix, tailoring offers just for you. These algorithms can be used for targeted upselling for online retailers or highly personalised offers.

  • An online wine retailer might recommend a particular type of wine, and nudge each buyer to a slightly higher price
  • A travel agent might tailor offers specifically for the demographics of each customer
  • A footy club might generate specific ticket offers to match the likely buying habits of each member

We also believe that this kind of technology has huge potential to increase workforce productivity by improving communication within a company:

  • Understanding what information each unique person needs in order to do their job well
  • Delivering that information at a time and in a way which is most effective for them
  • Using machine learning, to refine and improve that communications over time

As with any machine learning system, the more data you have, the more you can refine the model, and the more you can personalise.

The outcomes? Learning that sticks. New staff onboarded and quickly. Better adoption of new technologies...and your interactions with each employee more focussed and as helpful as possible.

If you’re still doing internal comms the old way - one-size-fits-all - now’s the time to chat.

And if you’re interested in reading more about the five patterns of AI, check out our recent blog: Behind the Five Types of AI.


Behind the five types of AI

We’re extremely proud of our track record of successfully deploying artificial intelligence and machine learning models to enterprise customers.

But artificial intelligence (AI) and machine learning (ML) are pretty broad fields. And we know you probably need a little more background on what we actually ‘do’.

So - here is a rundown of the five types of AI, or patterns, our work generally falls into - and how each pattern can bring benefits for your business.

Red Marble’s five types of AI

1. Prediction

Artificial Intelligence can give you the ability to predict the future! Or rather, it can predict the probability of a certain event occurring in the future, based on historical data.

Here are a few examples:

  • Forecasting conversion rate for an online retailer
  • Alerting when a machine is likely to be due for maintenance
  • Determining whether a member of a footy team is likely to sign up again next year
  • Predicting the likely valuation of a particular commercial property

To generate business value from prediction, you need to act on the information quickly - often in real-time. That combination of the ‘nugget of gold’ insight along with the right advice on how to apply it is what will bring big benefits for your business.

When we’re working with a client, our machine learning engineers and data scientists partner closely with subject matter experts who know the business well. We’re a big believer in these partnerships - we know AI inside out, but to properly capitalise on opportunities in different industries, we like to surround ourselves with diverse skills and experience.

 

2. Recognition

AI can recognise humans, objects or situations in a scene or recognise patterns in data. Businesses can use this in many ways.

Here are a few examples:

  • Recognising heavy machinery on a pedestrian area of a work site
  • Analysing whether a pizza has enough toppings as it comes out of the oven
  • Checking whether your staff are wearing masks during COVID lockdown

Businesses can use AI recognition to maintain the processes that make them great. Quality assurance teams can make sure each product heading out the door is at the right standard, and HR teams can use AI to ensure the safety of their teams.

 

3. Hyper-personalisation

This is one of the AI patterns we believe will generate tremendous value for organisations.

Hyper-personalisation is where we use machine learning across different data sources to build a unique profile of an individual. This means we can generate actions and content specific to that individual - such as personalized marketing or offers.

We’re all familiar with hyper-personalisation in one way or another. The most common example is your Netflix queue or the recommendations you get when you log into Amazon. This kind of engine - technically a set of algorithms - is hugely valuable for companies for engaging both their employees and their customers.

A few possibilities:

  • Offering specific holiday offers for families based on their demographics and location
  • Selling slightly more expensive wine to a regular online wine customer
  • Unique insurance offers based on a customer’s history, location or family status
  • Sending each employee a dynamic onboarding email based on their department, skills or interests

The possibilities of hyper-personalisation are really exciting for businesses. It facilitates one-to-one marketing, or unique learning paths for every employee. We’ve seen its power already, but we think it’s just getting started.

 

4. Conversation and language

This pattern of AI allows humans to interact with software using words - either by speech or text. The AI understands and infers intent and meaning from those words.

We believe that, in time, natural language will be the ultimate user interface. Just say or type what you want and the software understands. We do this already with Google Home or Alexa, but we see the application of voice assistants rapidly expanding into workplaces - especially as AI’s ability to infer meaning develops.

A few ways conversation and language AI could be used:

  • Recording, transcribing and uploading the CEO’s shareholder message
  • Analyzing work notes to see if any indicate a potential contractual claim
  • Listening to sales calls to infer the percentage probability of closing a deal

AI has the ability to change how people work within corporate environments - both helping employees save time and be more efficient, and by adding value at crucial points.

 

5. Outlier detection

This is the ability to detect unusual activity and behaviour outside of the norm. 

Anomaly detection is widely used in fraud prevention and cyber-security or to detect specific events from sensor data. It’s also increasingly used in medicine and in industrial environments.

A few examples:

  • Spotting cancers and anomalies in medical scans
  • Detecting when a machine is not running smoothly by its acoustic signature

This kind of detection is a powerful defence and protection against negative unforeseen events.

 

How can you leverage different types of artificial intelligence?

When people wonder what AI is, or what do we actually do in a machine learning project, the answer is usually one of these five patterns – prediction, recognition, hyper-personalisation, conversation and language or outlier detection. 

These forms of machine learning are already able to create value within your organisation. We encourage you to think about how your business works - would you benefit from being able to predict the future? Spot anomalies? Speaking to each customer or prospect individually?

To find out more about which types of AI can work for your business, get in touch.


How can AI help businesses better recover from COVID-19?

For now, at least, it seems Australia has successfully “flattened the curve”. 

We’re starting to emerge from lockdown - and thankfully, from a health perspective Australia seems to have been spared some of the horrors of other countries. With family living in the UK, France and New York, I’m acutely aware that the impact of COVID-19 has varied widely around the world.

From an economic perspective, COVID-19 has affected many industries – travel, retail, hospitality, entertainment and higher education to name a few. On the whole, technology companies have fared a little better as they navigate the current environment, but still face unique challenges as they establish their own ‘new normal’.

During this period, the importance of big data and trusted analytics has never been more vital. Artificial Intelligence has been in the news lately for a number of reasons, including the early detection of the pandemic and frontline resourcing and vaccine research.

How can we apply AI to help businesses recover from COVID-19?

We have conversations every day exploring the possibilities of AI with existing and potential clients. The challenges that organisations are facing right across the globe don’t have a playbook and navigating the next few months will undoubtedly be complex.

Our work generally falls into one or more of the following five areas:

     

  1. Prediction
    Being able to forecast future behaviour based on past experience, for example forecasting conversion rate for an online retailer or when predicting when a machine is likely to require maintenance.

 

  1. Recognition
    The ability to recognise people, objects or situations in a scene, often using computer vision technology, or recognizing patterns in data. This could include showing social distance between people or recognizing heavy machinery on a pedestrian area of a work site.

 

  1. Conversation & Language
    Enabling interaction between user and software using either voice or text. It requires understanding of a user’s intent and the ability to understand language and context. The equivalent of Siri or Alexa with the specific knowledge of a company’s “corporate dictionary” of acronyms, systems and processes.

 

  1. Hyper-personalisation
    Using machine learning across different data sources to build a unique profile of an individual and act upon this information with targeted marketing messages, personalised offers and unique discounts. It’s the ability to treat everyone as an individual rather than as part of a generalized market segment.

 

  1.  Outlier Detection
    This is the ability to detect unusual activity, patterns or behaviour which are ‘outside of the norm’ based on the available data.

 

Applying our AI methods to COVID recovery

Which of the above are most likely to be beneficial as companies return from lockdown? We’re seeing clear trends in the market around:

     

  • Recognition of people and situations
  • Hyper-personalised marketing
  • Prediction

 

The main focus for most companies is on returning to work and opening up to staff and customers safely. This requires a good understanding of the challenges within the physical environment.

Computer vision technology can be applied using existing security cameras for functions such as counting people in an area. Algorithms can detect people proximity issues, which can lead to either real-time alerts or heat maps to show flow of ‘traffic’. With some specialist equipment, heat-sensing technology can be applied to see temperature outliers within a group of staff or customers which might indicate a fever.

Companies are also looking at doing more with less. Hyper-personalised marketing can be applied to target product offerings to customers, re-engaging them without adding significant cost. Similar personalisation techniques are being applied within an enterprise to improve internal communications and  improve technology adoption.

Prediction using machine learning is likely to face challenges in the short term. Most models make certain inherent assumptions based on the data used to train the model, which may have seen large deviations from the norm.

Finally, automation technologies in general are increasingly being used to remove cost from operations. Though not typically “intelligent”, these are an effective way of automating many data processing tasks which can drive significant cost savings in the short term.

It’s been inspiring to see the innovation and ingenuity of so many businesses as they navigate this incredibly challenging period. It’s great to see that many businesses are looking at technology as a solution to the current obstacles.

Now more than ever, companies need projects to deliver value quickly. If you would like more details on how Red Marble can help, please get in touch.


Six key questions towards Ethical AI

As companies emerge from these tough times of COVID-19, many organisations are looking to take advantage of Artificial Intelligence (AI) to build efficiency and improve customer and staff experience. Along with the many benefits, there are also risks and liabilities to consider - particularly when it comes to inheriting bias.

It’s critical to deploy any artificial intelligence project with ‘Ethical AI’ as a key lens as we move into an equitable future.

AI might seem immune from the moral tangles and complicated contradictions of humans, but in reality it’s difficult to escape the biases that unknowingly sit in the data used to train the models. Not only does this create an ideological issue, it also affects the business outcomes and increases the risk of negative impact.

Ethical AI - six questions to ask

As the market and more importantly the influence of AI continues to build, we need Ethical AI to be our directional “North Star” to ensure we manage the inherited risk of bias within AI models.

At a minimum, we suggest asking and discussing six key questions throughout the process of procuring, designing and implementing AI projects. This may be a project sponsor asking a vendor, or a project manager asking their developers, or a CEO asking a CIO:

  • Which laws and regulations does the software need to adhere to? 
  • How does the solution design enable any decisions, or recommendations made, to be explained? 
  • Is there transparency about this use of AI with each stakeholder group? 
  • What biases may be inherent in your training data? How are you managing the risk of bias?
  • What decisions will the software make or inform? Are the decisions reversible? 
  • What’s your process around code reviews and how are you adopting software engineering best practice?

Leading the charge for responsible AI

Many governments, organisations and businesses, particularly in Australia, have voiced their support of the Ethical AI movement.

The Australian Government’s Department of Industry, Science, Energy and Resources has created a list of eight voluntary AI ethics principles; among them, “Throughout their lifecycle, AI systems should benefit individuals, society and the environment.” However, we feel that most companies need a more practical approach and find those principles a little too high level. 

Red Marble AI has an ethical AI playbook for companies looking to establish their capability quickly - please reach out for a copy. There are also a number of specialist ethical AI firms emerging, such as Dr Catriona Wallace’s new advisory which recently opened out of Sydney.

Ultimately, we expect companies will form an AI ethics advisory board or sub-committee, similar to existing risk and audit committees, to develop a framework for a company’s expectations around AI. But in the meantime, we believe that asking these six questions – and the discussions that ensue - will give most companies a strong head start.

It would be great to hear your thoughts on Ethical AI, and our six principles. How strict should we be? Do we need government legislation? Please get in touch here.