So it is that what used to be called 'middleware' — the hidden and often arcane field of enterprise application integration — is suddenly sexy again. Or at least, it's back in demand. But like many survivors from the client-server era, the technologies that worked then are no longer fit for purpose in the digital era. They were created at a time when 'integration' was something that was rarely needed and, once completed, seldom needed updating. It could safely be left to a skilled cadre of IT specialists.
Nothing could be further from today's digital world, in which connection is the lifeblood of success and change is constant. The number of connections and the frequency of updates mean that it's no longer viable for a small band of specialists to hand-craft every separate integration. The connections must be ready-made, and easy to deploy off-the-shelf. Thus we see integration vendors such as Dell Boomi, MuleSoft and SnapLogic bringing out low-code development tools, talking up an API-first strategy, and emphasizing self-service approach.
Intelligent workflow automation
Even these improvements may not be enough to bring enterprise application integration fully into the digital era. The founders of Workato, which last month revealed a $10 million series A round from its backers Storm Ventures, Salesforce and Workday, see a need for much more automation to deliver all the application connections enterprises will need. The latest Workato platform update, announced at the same time, uses intelligent automation to "consumerize integration," says Bhaskar Roy, head of growth. This makes it easy for business users to build automated workflows and insights that connect across multiple apps and data sources, he contends:
Integrations were built to last. The world has changed today. Integration has to be something that changes fast.
Integration can no longer be limited to IT. [The aim is] enabling many more people to build integrations, but still have IT have that oversight.
The Workato team have the credentials to win the trust of IT, even while they reinvent integration. The startup was founded by a core team that were involved in the early years of enterprise application pioneer Tibco, including Gautham Viswanathan, head of products. It already claims more than 21,000 organizations using its platform. Visswanathan tells me the time is ripe for Workato's new thinking:
For many of us it's the third or fourth product or company in this space.
There has never been a better time to have a product that can be very widely adopted. We feel this approach of combining power with access gives us the ability to make integration available to a much larger audience beyond just IT.
An example of what this means in practice is illustrated by Workato's integration of its Workbot intelligent agent technology into messaging platform Slack. It's easy to build a Workbot that brings alerts or information into a Slack channel from applications such as Salesforce, Zendesk, Jira, Workday or many others. The user can respond via the bot without having to open up the originating application. Workato has also linked its Workbot into other messaging platforms including Microsoft Teams, Hipchat, IBM Workspace and Cisco Spark.
This trend towards messaging acting as a conversational engagement layer with back-end apps that thus become 'headless' is one that's up-and-coming. It's being enabled by a combination of platforms like Workato that make it much easier to link bots to back-end applications, together with messaging platforms that are open to this kind of interaction. Using Workato not only makes the interaction more user-friendly, it also brings more oversight of what's happening in the message stream, says Viswanathan:
With modern chat products, these Slack bots give you the ability to provide good UIs.
Enterprise decision making is moving to Slack, and the approvals for those are moving to Slack. But there's absolutely no tracking. What Workato allows you to do is to build these so that there's an audit trail.
You can build these bots without having to write a single line of code.
Recipes for reuse
The way that Workato approaches integration is by thinking in terms of automation — this focuses attention on the process and the outcome it's intended to produce, rather than the integration per se. Each automation is defined as a recipe, comprising a trigger and a series of actions. The steps can be conditional and can call other recipes.
Workato encourages reuse, seeing the GitHub repository of shared code as a model to emulate for its recipes. "Its just like GitHub," says Roy, "you build your integration and you can choose to make it public." There are already 150,000 recipes available. In most cases, therefore, it's possible to adapt an existing recipe rather than having to build it completely from scratch.
The latest release introduces machine learning assistance. For example, when a user first picks up a recipe, the platform will highlight any parts that don't work, perhaps because it calls a field that doesn't exist in your application instance. As you continue building the recipe, the platform makes recommendations as you go along, based on its learning from what others have done in the past.
When deployed to production, a recipe runs in its own container to maximize execution speed, and the system dynamically scales based on load. It also "automates the automations," with exception handling to recover quickly from data or process errors such as API rate limits, network disconnect, API going down, or expired authentication tokens. It uses machine learning to automatically detect anomalous data traffic patterns too, and to prompt admins with suggested responses.
Workato says that just over half its users (51%) are in business roles such as sales operations, administrators, and so on, while just 23% are in IT roles. It also claims that 78% of users go live with their first integration project within a week. Adoption tends to spread rapidly within an organization, says Roy, and the product lends itself to a more iterative approach than traditional integration projects:
The rate at which IT is expected to add new integrations, it needs help from other parts of the company. Workato is a product IT can hand out to other departments to do their own automation. But IT still wants to be in the loop — we have a ton of governance that goes into it.
With our platform, you don't build the mother of all requirements. People just build it and incrementally add.
Workato's use of machine learning to augment the skills that users bring to the creation and operation of its integration recipes gives it an important differentiation against more established competitors. Latching on to Slack and other messaging platforms is also a smart move. At the same time, its founders have the credibility of longstanding familiarity with the concerns of IT when it comes to enterprise integration. We'll be keeping an eye on Workato and aim to dig into some customer stories in the coming months.