Boomi doubles down on AI and automation

Boomi showed off its artificial intelligence (AI) and automation capabilities at the recent Boomi World Tour 2023 in Sydney, where CEO Steve Lucas told a packed room about the changes that were needed to unleash the full potential of those technologies.

Take data, for example. Noting that while data is important and abundant, only a small fraction of it is used, so rather than seeing data as the new oil, the mantra should be “data is the new sand”, focusing on the parts that really matter, he said.

The digital footprint of organisations is also increasingly fragmented. Citing various sources, Lucas said the average enterprise has 324 cloud apps, 79% of organisations have over 100 data sources (30% have more than 1,000), and yet only 19% of boards feel their organisations have made progress towards their digital transformation goals.

Furthermore, between two-thirds and almost all businesses responding to surveys said they had lost or missed out on business due to integration issues. And integration, of course, is Boomi’s bread and butter. Large language models (LLMs) are the current wave in IT, he said, and because this lets “computers talk like humans instead of humans talking like computers”, one consequence is that “good programming today starts with a good generative prompt, not code”.

When it comes to things like service requests, “there will be a time when you prefer to talk to the machine” rather than repeatedly pressing zero on the phone keypad in the hope of being transferred to a human operator.

One of the Boomi platform’s latest capabilities is Boomi AI. Based on an LLM with knowledge of the 200 million integration patterns created by Boomi users, plus various other models of unspecified types, it became generally available to Boomi customers in the past few weeks.

The first phase is Boomi GPT, an autonomous design tool that can respond to prompts such as “create the process to connect AutoTask to Salesforce” by requesting additional details of what is required, and then generating the integration.

Boomi GPT is capable of building integrations, application programming interfaces (APIs) and master data management models. The company plans to develop Boomi AI further to provide autonomous management and autonomous orchestration, and Lucas said it would soon be possible to ask questions such as why invoices had not been processed.

A task automation service is being developed and is currently expected to be released in the first quarter of 2024, chief product and technology officer Ed Macosky told Computer Weekly.

This will democratise the creation of automations, giving users the choice of specifying what they want through either natural language descriptions or assembling tiles in a visual interface. At the same time, IT departments will be able to enforce governance and compliance issues through mechanisms such as role-based access controls applied at run time. And if they see that dozens of people have created similar integrations, they can produce an enterprise-grade version and make that available to everyone.

Customers have also asked for the automatic production of documentation for AI-generated components, he said. Also on the roadmap are enhancements to Boomi GPT, special AI connectors, code generation, and autonomous management and orchestration.

AI features in the Boomi Standard Edition are Suggest, a data mapping assistant; Resolve, which provides possible solutions to document or process errors; and Quick Start, Boomi’s no-code environment.

The Pro AI Edition adds Boomi GPT, and the Enterprise AI Edition will include everything from the other editions plus the forthcoming features outlined above.

Also in the pipeline is Task Automation, a mobile-enabled tool that will allow users to create their own automations from spoken or typed natural language descriptions.

Macosky gave an example of creating a Slack or SMS message from an incoming email, and pointed out that Task Automation will operate under the governance of IT departments. This could, for instance, prevent personal identifiable information from being sent outside the organisation.

“Productivity is the keyword,” said David Irecki, Boomi’s director of solution consulting in Asia-Pacific and Japan, adding that Boomi AI can be used to reduce costs and redeploy integrations to high-value projects.

At the other extreme, because it allows organisations to add their own transformers, “Boomi can be at the heart of your AI strategy,” Lucas observed.

There is lots of interest in AI in Australia and New Zealand (ANZ), Irecki pointed out. At the coalface, some are keen on the technology as they see it helping them do their jobs in the “co-pilot” model, although others are scared it will make them redundant.

But “integration is ubiquitous” – there are always things to integrate and automate, but inflation is still high – adding to cost pressures – and the skills shortage still exists. The challenge is to use technology to free up skilled workers to concentrate on the higher-value aspects of their jobs, and also to automate tasks such as data entry while preserving the process knowledge held by those currently doing those tasks.

Australian Red Cross CIO Brett Wilson said that with Boomi AI, his organisation will not need as many technical integration specialists, and that will make room for more business analysts who can deliver more value than simply building connections between applications.

He also described the part Boomi played in making possible the telethon in support of flood victims in New South Wales and Queensland, saying “Boomi was one of the technology partners that made it a success”.