Natural language technology, no code and active learning will help intelligent automation leaders scale their hyperautomation plans in 2022.
After a year of rapid and stunning adaptation to the pandemic, 2021 was a time of consolidation for intelligent automation leaders. 2022 will be no less exciting but the tenor will be very different. Now that the foundations of new AI and automation technologies have been secured, digital transformation leaders will seek to capitalise on them. To do that, they will need scale.
2022 will be a year of profound innovation as change agents explore new ways to achieve hyperautomation - scaling the benefits of AI and automation to every corner of the enterprise.
Here are the most important trends for intelligent automation leaders to follow in 2022:
Conversational AI and natural language technology see rapid adoption
For the longest time, intelligent automation has focused on a restricted stack of technologies, with robotic process automation (RPA) at the centre. In 2022, the RPA market is expected to continue growing - increasing by 53% to $3.17 billion. Yet the realisation will also sink in that RPA alone cannot deliver what digital transformation agents had envisioned.
For existing adopters, we’re already reaching the end of the road for what RPA alone can achieve. Restricted to simple, repetitive processes, RPA will struggle to make an impression on the most wasteful and inefficient processes that can’t be automated with bots. Finding new simplistic processes to automate through process mining and process mapping can only get you so far.
That’s why intelligent automation leaders will widen their toolset in 2022 to include natural language technologies and emerging categories like Conversational Data Intelligence. We have already seen the green shoots of this process in 2021 with the successful Series E funding of Sales Conversation Intelligence platform Gong, and ZoomInfo’s acquisition of Chorus.ai. An impressive 67% of business services leaders have also been putting investment aside for natural language processing (NLP) to extract data from emails.
The benefit of these natural language technologies is their ability to extract precious insight from previously untapped sources. Leaders will gain new-found business intelligence by performing Communications Mining on the digital conversations of their employees. Meanwhile, the same technology will help employees understand customers like never before, analysing their contact and feedback at speed and scale.
However, there’s also an important automation component. By creating structured data from business communications, Conversational Data Intelligence will open up whole new pathways for automation systems. For the first time, businesses will be able to perform the end-to-end automation of complex, conversation-based workflows - from answering a customer’s frequently asked question, to triaging an incorrectly sent email to the right team.
Intelligent automation leaders can expect considerable savings, while also giving precious time back to employees to focus on the most valuable and rewarding work. Those who get there first are sure to see a major service efficiency and competitive advantage over the competition.
Research and analyst houses are really starting to recognise the value of natural language technologies. Download your free copy of Gartner’s new Cool Vendors in Conversational and NLT report for more analysis of these trends.
No-Code development goes into overdrive
2022 will be a huge year for the citizen developer. The experience of the pandemic has shown that business-as-usual software development can no longer meet growing demand. This is especially true when it comes to AI. The pool of data scientists and engineers with the experience and skills needed to develop AI applications has always been small. Yet, development talent has never been in higher demand given ongoing disruption to markets and the digital ecosystem.
In 2022, amid a severe shortage of AI talent, we’ll see businesses embracing low-code and no-code tools that empower every employee to help create the latest services, updates and platform improvements. More than three-quarters (77%) of enterprises are adopting low and no-code, with 45% of IT leaders seeing faster and more collaborative development as a result.
These tools have already created a new breed of AI ‘citizen developers’ in the enterprise. These are staff who work on developing the latest models but whose primary job function remains outside the field of AI and machine learning. This emerging cohort will become pervasive in 2022 with Gartner predicting that by 2024 80% of technology products and services will be built by those who are not technology professionals.
If 2021 was the year no-code achieved unprecedented uptake, 2022 will be the year businesses formalise its use and capitalise on its potential. AI and machine learning are the great levellers of business success, but upfront costs and skill shortages have long put a brake on adoption. No-code will be key in spreading AI and machine learning across the enterprise and achieving hyperautomation.
Employee Experience and Total Experience are the new focus of intelligent automation
When the C-suite considers the benefits of automation, cost savings and efficiencies will usually top the list. However, as automation becomes more widespread, business understanding of the concept is becoming more nuanced. Not only can intelligent automation cut costs and create value - it also enriches the working lives of employees.
This came into sharp relief in 2021. The disruption of the pandemic had a major impact on employment and retention around the world. In the US, the Great Resignation gathered speed as workers gave up their jobs to focus on activities they saw as more fulfilling. Meanwhile, in the UK a record 400,000 people switched jobs in the third quarter of the year, and 60% of managers say it is now harder to recruit than before COVID-19.
The talent shortage matters to digital transformation leaders because it’s going to heavily inform their priorities in 2022. The C-suite will push intelligent automation leaders to help make the employee experience more satisfying and rewarding. To stave off further resignations and customer churn, change agents will spend more time improving the total experience - the combined brand experiences of employees, users and customers.
This is actually the perfect fit for intelligent automation. When leaders combine automation and AI technologies to build more efficient and intelligent workflows, they improve everyone’s experience of the company. Intelligent automation removes the mundane, repetitive tasks employees don’t enjoy, freeing them to focus on their most valuable and rewarding work. A great example of this is how Communications Mining and RPA work together to understand and automate mundane, repetitive comms work - like forwarding an email to the right team, or processing a transactional request.
By automating these manual workflows, change agents also improve service speed and delivery - creating a more satisfying customer experience as well.
Active Learning will make AI scalable
Increased adoption in a wider range of AI technologies is a given while intelligent automation leaders seek to extract more value from their automation initiatives. But, as well embracing no-code and low-code to speed up AI development, they’ll also have to explore new ways of accelerating production.
One of the greatest limiting factors on AI adoption has long been the time and investment needed to train it. For example, models trained using a popular cloud NLP solution like Google AutoML can take weeks or even months before they are actually ready to start producing value. Many projects are scrapped before they can even hit the ground due to what the business perceives as a lack of progress in slow training times.
This is a considerable problem, for it doesn’t give AI the chance to scale in your organisation. A great deal of resource and investment goes into intelligent automation, and for it to return strong ROI it needs to be deployed on a large scale. But the time needed to train or update a model for a new use case can really scupper this project.
That’s why in 2022 intelligent automation leaders will be pushed to explore new and more efficient model training methods for enterprise AI. We expect that many will hone in on Active Learning platforms that promote collaboration between human subject matter experts (SME) and machines.
Traditionally, training machine learning models requires thousands of examples - each labouriously labelled by employees. Active Learning significantly reduces this requirement by only querying users with the most valuable data points.
In practice, Active Learning requires far less manual labeling from employees and creates highly accurate models in less time. In our own comparison between Re:infer and Google AutoML, we found that Active Learning contributed to model training that was 200-times faster than the alternative. Read our full comparison.
