Automation Projects Could Drain Your Finances. Here’s How to Get It Right

Automation has become the latest buzzword in the technology industry, but kickstarting an automation project and turning it into a success isn’t an easy task. Liviu Dragan, the CEO of conversational AI chatbot provider DRUID, offers interesting insights into why most automation projects fail to take off and how developers can turn it into a success story.

Unlike what it was a decade ago, automation has permeated every industry and enables many use cases. Fortune Business Insights predictsOpens a new window that the global industrial automation market will reach $326 billion in 2027 and will grow at a CAGR of 8.9% in the period.  

Gartner also foresees significant growth in the adoption of automation and hyperautomation technologies worldwide over the next few years. It predicts that industrial production will be automated to such an extent that by 2025, customers will be the first humans to touch more than 20%Opens a new window of the products and produce in the world. It also predicts that 50% of enterprises will devise AI orchestration platforms to operationalize AI by the same year. By 2024, organizations will lower operational costs by 30% by combiningOpens a new window hyperautomation technologies with redesigned operational processes.

Such drastic changes can only be brought about by significant investment in automation technologies and the availability of highly skilled developers. Nasdaq terms automation as “The Next Big ThingOpens a new window ,” noting that while declining fertility rates and a dwindling labor force are hampering productivity growth in the U.S., “automation could be the productivity boost the U.S. needs.” “With the development of cloud-based platforms, corporate automation tools are now accessible to even the smallest companies, and we’ve seen the pandemic accelerate its adoption,” it said.

At the core of this massive shift are small automation projects that, leveraging machine learning, Big Data, and advanced analytics, are continuously being refined to replace inefficient processes and bring about new inventions that cater to changing customer needs. The advent of 3D printing, assistive robotics, drone deliveries, AI chatbots, and farm automation are a few examples. 

Learn More: Is Intelligent Automation an Alternative to RPA? Experts Weigh In

Why Is It So Difficult to Make an Automation Project Work?

An automation project often starts with a group of developers determining the project’s scope and objectives, devising a software development process, continuously testing their DevOps approach for inefficiencies, testing their software for flaws, and measuring performance at each stage of the development cycle. However, many stumbling blocks are beyond the control of developers that threaten to destroy the project at any stage.

Cost: Cost is a major determinant of whether an automation project is worth investing in. One of the first tasks of an automation engineer is to determine how much would it cost to hire skilled professionals and high-performance machines, how much would it cost to continuously refine and upgrade the software, and the cost of taking the product to the market. Failure to accurately estimate these costs could lead to cost overruns and loss of interest from investors. 

Expectations: Each automation project carries with it huge expectations in terms of increasing efficiencies, reducing costs, improving time-to-market, streamlining operations, etc. Often, there is a gap between what the Board thinks a project can do and what it really is capable of doing. Unless automation engineers are able to clearly define and communicate the scope and limitations of their projects right at the outset, the management can quickly lose interest and move on to other projects.

Developer Skills: Before kickstarting an automation project, it is essential to build a core team consisting of automation engineers and developers who have the right skills and experience to ensure the project goes past the finish line. It is often difficult for smaller firms and startups to afford experienced workers who may fit the bill. However, this issue can be overcome if the firm can generate sufficient investments.

Poor Planning: Planning an automation project not only involves the setting up of a watertight development process and project timelines, but it also involves calculating the costs, the requirement of people and systems, estimating the capabilities of developers, accurately understanding the cost of the project, and enabling a consensus between stakeholders. Failing to calculate or evaluate any of these criteria correctly can sink a project very quickly.

Jouko AhvenainenOpens a new window , the co-founder and COO of Robocorp, saysOpens a new window that the bleak success of automation projects (11 out of 12 of them fail to take off) complies with the time-tested notion of no war plans surviving contact with the enemy. He narrows down the reasons behind the high failure rate to strategic problems, problems with the business model, and implementation and technology problems.

According to Ahvenainen, automating the independent routines of individual employees is not enough to leverage the real value of RPA. To make their processes more effective, businesses must focus on projects that enable different parts of an organization to function efficiently as a whole. At the same time, organizations often use low-code tools in their projects but quickly realize that they are not robust or scalable enough to support demanding requirements.

“Low-code works for some simple tasks of individual employees, like Excel macros work. No-code also works for some purposes, like making web sites from standard components. But if you really want to make a professional software implementation, you need professional tools and developers. Why otherwise world leading technology companies pay $250,000 or much more for professional developers?” he asks.

Learn More: How Purpose-Built Intelligent Automation Solves the Shortcomings of RPA

How Can You Get It Right?

To understand how automation projects can be salvaged or run successfully, there is no better guide than someone who has steered many projects, turned them into success stories, and minutely understands how automation works. Toolbox spoke to Liviu DraganOpens a new window , the CEO of DRUID AIOpens a new window , a provider of an AI-powered, no-code, chatbot authoring platform that allows citizen developers to design, develop and deploy interactions between employees, customers, partners and enterprise systems through omnichannel text and voice conversations.

According to Druid AI, AI-driven chatbots are among the most impactful inventions driven by a combination of machine learning, natural language processing, RPA, advanced analytics, and process mining. By 2023, the use of AI-driven chatbots will help banks eliminate $7.3 billion in operational costs and help the retail industry save $439 billion by optimizing processes with chatbots.

The Best Way to Begin

According to Dragan, building the right team at the outset is a significant factor that determines the success of an automation project. “The most important lesson that I learned is to have the right people around you, whether they are advisors, venture people, employees, and friends. Our growth is so accelerated that I think we would not have made it unless I had people I could rely on. I focused a lot and involved myself in the recruitment of the core team. I knew exactly who I wanted to have the DRUID journey started,” he said.

Another major determinant understands the product and its need for constant updates. “Our product is alive and needs constant training and updates. This process means that, from time to time, we had to stop, reevaluate the decisions made and then start over. Being the CEO of a start-up feels like being on a rollercoaster daily, juggling clients, meetings, partners, solving problems continuously while caring about the employees. The result is innovative and exciting, even though the ride is challenging at times,” Dragan said.

Solving the Key Challenges

Dragan says the top challenges a startup faces when initiating a project are having the right people in place and finalizing a pricing strategy. “Among the first people that I recruited was an experienced architect because a product without a robust architectural base cannot exist. If the start is faulty, you will undoubtedly face serious technological and strategic problems as you progress. 

“Another challenge we had to tackle was the pricing strategies. In the beginning, we had a price scheme that was based on the idea of simplicity, but in time it proved to be somewhat confusing for both our partners and clients. Consequently, we had to change it and do it quickly.”

Even if firms manage to overcome these challenges, Dragan cautions that they should never forget about the importance of pilots. “The pilot, or POCs, should be done in a controlled environment that demonstrates the powerful impact of automation and, from that point on, plan its expansion. Especially for the international markets, where the competition is aggressive, we are using POCs to prove the superiority of our product in comparison with other platforms.”

As for training virtual assistants, Dragan says that it is no different from training a human employee and requires time and investment. “Start with a clear strategy and set up key metrics to evaluate the success of the initiative. Mobilize the right team (IT and subject matter experts) to build a Minimal Viable Product to start the learning experience and demonstrate the usability and impact of a small-scale. Further, expand the initiative at the organizational level, communicate internally to embrace the benefits, and don’t forget to give a personality to every new chatbot.”

Learn More: Top 8 Open Source Tools To Up Your IT Automation and Event Correlation Game

How to Boost Your Skills And Ace an Automation Project

Following are some characteristics, according to Dragan, that are essential for data scientists and data engineers to succeed when working on automation projects:

1. Choose any language but master it – When it comes to building and testing algorithms, the language of choice is less important than competence in that language. Choose one that lowers iteration time or allows you to test an idea in the shortest amount of time.
2. Encourage curiosity and questioning by reading books and using internet resources such as Stack Exchange, subreddits, podcasts, and YouTube about ML and AI.
3. Develop your domain expertise – Before diving into machine learning and artificial intelligence essentials, it’s vital first to grasp the industry’s current developments.
4. Master Technologies & Programming Skills. To get a leg up in this area, you’ll need a lot of project experience. Pay attention to deployment as well, even if you’re concentrating on development.
5. Human Intuition – While we should rely on data-driven insights, human intuition will continue to play an important role. This is because making decisions will get more difficult as the world becomes more complex. In this situation, combining careful investigation with intuition is a solid technique to follow.

Do you think poor planning is the main reason behind the failure of automation projects? Comment below or tell us on LinkedInOpens a new window , TwitterOpens a new window , or FacebookOpens a new window . We would love to hear from you!