The AI Squared Enterprise: AI, Automation, Integration

4 minute read

Hg's 2024 Digital Summit in Amsterdam brought together over 120 technical leaders to explore emerging use cases and the latest advances in AI. During the summit, Hg's James Cope, Business Systems and Transformation Specialist, speaks with Markus Zirn, Chief Strategy Officer, Workato, on all things AI.

Generative AI is expected to be complementary to, rather than reduce the need for, business automation and integration, supporting further industry innovation over the coming years.

Integration is the plumbing behind automation, helping companies to unlock data in silos and make it more widely available in business processes. Today, the ability to fuse Gen AI services together with automation to move data sets from system to system, presents a big opportunity. For example, this could relate to data mapping between IT systems, with Gen AI being utilized to help with the design of automations, as opposed to supplanting them altogether, in turn increasing business process efficiency.

High complexity low volume use cases

While there is an acknowledgement within the software industry that Gen AI presents myriad opportunities, there is also a degree of nervousness and risk among business owners about becoming irrelevant.

This was a point raised by Hg’s James Cope, Business Systems and Transformation Specialist, during a discussion with Markus Zirn, Chief Strategy Officer, Workato, at the Hg Digital Summit 2024.

In Zirn’s view, much of what has been automated in the past has tended to be in high volume, low complexity use cases; as such, many of today’s business integrations are still fairly rudimentary. Rather than be viewed as a risk per se, Gen AI services could present a huge opportunity for software companies, leveraging them for high complexity, low volume use cases that increase the overall quality of existing business automations.

One of the barriers facing software companies is that they have a lot of valuable data locked
away in different systems. However, unlocking that data can involve investing significant time and effort.

Although it is far from a panacea, Gen AI could help accelerate how people think about utilizing that unlocked data in a more accessible way. Not only could it help access data, it could also make it easier for software companies to leverage that data.

Enterprise automation and Gen AI are complementary, in terms of unlocking valuable data and creating easier integration between systems. Pointing Gen AI at one system and telling another what the data mapping will be can improve business automation, where the Gen AI is able to flag exceptions for humans to take action and decide on next steps.

“We express automations as recipes (basically information flows) that connect to different systems,” Zirn remarked. He said that building end point connectors used to be coding exercises. “Now we can use a co-pilot, where we point it at the API and it automatically creates the end point connector. If I get my data mapping suggested as well, that is very helpful.”

An explosion of connectivity

As the AI revolution accelerates, it has the potential to create a vast array of new technology companies. The net result of this could mean that companies will face having to connect to even more systems, while the range of data models becomes increasingly disparate. This could lead to the formation of new, innovative companies to address the connectivity challenge.

Data is the lifeblood of AI. The better the quality of the data, the more reliable AI is. The next logical step is to automate machine to machine and take humans out of the equation. Previously, the focus was on using structured data only for business automation and integration.

Now, companies can use Gen AI to handle unstructured data to cope with wider data demands.

Product development is another area of potential benefit. By integrating quantitative and qualitative data, software companies will be able to better understand not only how their customers are using a particular product, but also how they feel about using the product. In that regard, it is not just about the quality of data at the core of a system, but the ability to use that data to drive insights on future product design and innovation.

From Author to Editor

Generative AI offers strong potential to enhance business integrations with its ability to handle unstructured data. Not that this will discount the need for human intervention entirely. People will still be required to work on data gaps and seek out additional information. While this might only be 20% of the time, it is likely to present a permanent barrier to moving towards truly autonomous processing.

The traditional mindset for automation has tended to apply to low complexity, high volume tasks (think straight-through processing). Gen AI provides an opportunity to compliment this more robotic automation by figuring out and flagging anomalies.

This change has been described by some within the software industry as moving from being an author to an editor. A given business automation runs as intended, but once gaps or exceptions are identified, the human steps in as required, without having to do all the heavy lifting. This reduces friction, and increases the speed and efficiency of automations.

"Sometimes the best automations bring humans in at the right time. It might feel a bit counterintuitive but if you use Gen AI for anomalies, that’s the best thing you can do.”

As people look for ways to communicate with computers to get better outcomes, their interactions will change and will, in turn, impact how people change the experience in processes and in software.

Process redesign

This is still the first innings as companies begin to consider Gen AI for higher complexity use cases. The software industry has barely scratched the surface of what might be possible. Use cases have tended to concentrate on certain areas, such as sales and operations. As companies develop a deeper understanding of the technology, 2024 could see more business innovation, and a greater willingness to do more experiments to figure out where the value is.

To do so will also require a change of mindset.

In the enterprise automation space, the real innovation is likely going to come from those willing to determine whether their existing processes should change – or whether new processes should be designed – to better take advantage of AI. Especially given that new classes of unstructured data will support innovation, opening up numerous pathways to improve integration and automation.

As the proliferation of AI systems grows, companies will be required to orchestrate the different AI models they adopt to fulfill their end to end business processes.

Share this article