IBM Watson Orchestrate: a dramatic step forward but will it work?
Watson Orchestrate is part of IBM’s larger automation strategy to help users be more productive by automating tasks. In IBM’s own words:
Your own personal AI, in the tools you already use.
Watson Orchestrate gives you interactive AI–in tools like email and Slack–to increase your productivity. This isn’t a static bot programmed by IT. You initiate work in natural language, and Watson Orchestrate uses a powerful AI engine to combine pre-packaged skills, on-the-fly and in-context, based on organizational knowledge and your prior interactions.
It’s much more than a virtual assistant but the end-user empowerment opportunities stand out to me as its most salient results that business executives should pay attention to. Watson Orchestrate applies AI to achieve business results.
Back in the typical enterprise, most are pursuing AI by:
- Waiting for the vendors of their important enterprise applications (or those vendors’ more innovative partners) to deliver significant, AI-derived value in new versions of those applications
- Investing in virtual assistants for specific use cases (typically upgrading interactive voice response technologies or scripted, speech-enabled decision-tree walkers.)
The first approach is several times more prevalent than the second.
IBM’s move has the potential to disrupt both of these patterns.
Working a high wire without a net?
Two fundamental challenges for this IBM initiative to survive and thrive:
- IBM must be clear in both intent and execution on support for a much broader range of enterprise applications than it has announced. More enterprise applications have to be added to the Watson Orchestrate portfolio quickly enough to maintain the attention of customer executives whose key applications aren’t yet on the Watson Orchestrate supported list.
- IBM is going to have to risk going against tight IT internal control to make the Watson Orchestrate really blossom.
Conversational AI Platforms (CAIP), which I have been focused on for more than five years, are the foundation for entirely new ways of working.
IBM Watson Orchestrate uses CAIP to empower (and advise) users – ‘mere mortals’ – to tell their IBM-provided virtual assistant how to integrate the flow of processes and work across different enterprise applications.
For now, the IBM product only integrates across Salesforce (including Slack), SAP, and Workday applications, as well as various Email and Calendaring tools. It does it in a way that will feel to many enterprise users like a miracle.
Instead of waiting years for the IT organization to design, develop and deliver cross-system integrations, IBM claims users can talk to their virtual assistant and ask Watson Orchestrate to do-what-I-mean (DWIM.)
DWIM is a term near and dear to my heart. Back in the 1980s, when I was product marketing manager for Digital Equipment Corporation’s ALL-IN-1 Office Automation Suite, we dreamt of having a DWIM button on terminals.
Today, in 2021, IBM is beginning to deliver DWIM. “Tell me what you want or let me watch, and I’ll make suggestions.”
The Innovator’s Conundrum – which branch to take?
Success requires IT buys into this multiplication of integrators and integration points and allows users to access and use these special powers.
- The risk for IBM and others following this same no-code “intelligent” virtual-assistant buddy model is that IT, their primary buying center, might dramatically delay uptake.
- The risk for vendors who aren’t following a path similar to what IBM is doing is that enterprises will “see the light” and realize they can’t hold back the end-user-empowerment tide.
Is this magic?
It sounds like it. How many different enterprise applications from how many other suppliers can they integrate? When will they support Oracle’s applications? ServiceNow? SAS? Atlassian? The other top 100 software application providers? What’s the roadmap and timeline?
It’s more than purchased applications. How do they discover all the custom interfaces that IT-developed applications and extensions have? Integration requires a full understanding and mapping of interfaces and semantics of APIs and underlying functions. How can they understand your applications and data if you can’t either?
Bet on it:
- Virtual assistants are evolving to be much more than chatbots that replace “interactive voice response” systems.
- Conversational AI Platforms (CAIP) will “rule the roost” observing and predicting, empowering and guiding users to create their own custom individualized work environment.
- IBM has taken a very early lead staking out this territory, but there are many opportunities for other vendors as well.
What to do next:
- Cheer on the vision behind Watson Orchestrate. Let’s see what they can execute.
- Watch for early successes in your industry and similar announcements from other CAIP vendors or service providers.
- Start a canvas of CAIP vendors in addition to IBM. Go beyond “all the usual suspects” and include small and midsize suppliers because this is an early-stage market.
- Game out the impact and pursue the innovation ROI hotspots:
- Identify the specific applications your organization is dependent on.
- Use internal wisdom-of-crowds and other techniques to surface novel, high-value use cases of composites created by integrating across applications.
- Test those use cases while paying particular attention to cases that provide quick returns at low cost.
- Factor changes in “technical debt” in your cost analysis.
- Start executing.
Reach out to me. I can help.
IBM Quote reference: https://www.ibm.com/cloud/automation/watson-orchestrate/