Troy
Troy Bjerke
Senior Design Lead — IBM
Navigation

BJERKE / DSGN

Senior Design Lead · IBM

← All Projects
watsonx Orchestrate Unified Automation Builder

IBM · 2023

watsonx Orchestrate Unified Automation Builder

Design Lead

EnterpriseAIAutomationProduct DesignConsolidation

This project is the origin of the canvas-proliferation problem that Troy later resolved on Agent Builder. By focusing on connecting tissue rather than individual product redesigns, the team shipped a working consolidation — but created multiple canvas/visual building experiences with overlapping patterns. Recognizing that problem early shaped the design strategy for the Agent Builder project that followed.

Challenge

The Problem

What we were solving

IBM had three separate automation products — IBM Business Automation Workflow, the old watsonx Orchestrate Skill Builder / Conversational Actions, and Automated Decision Service — each with its own UX patterns, mental models, and interaction paradigms. Users moving between them hit constant friction. They needed to live together inside watsonx Orchestrate as a unified experience.

Design goal

Consolidate three distinct automation products into one coherent experience within watsonx Orchestrate, without wholesale redesigning each individual product — instead focusing on the connecting tissue that makes them work together.

Constraints

  1. ·

    Could not fully redesign each individual product — too much existing user investment

  2. ·

    Three teams with different technical constraints and roadmap timelines

  3. ·

    Must maintain backward compatibility for existing users of each product

  4. ·

    8-designer team across multiple geographies requiring strong design coordination

How we learned

Research

Findings

Research showed that users weren't moving fluidly between the three products — they were siloed by the tool they'd learned first. The primary insight was that the mental model problem was as significant as the UI problem: users needed a unified way to think about automation, not just a unified interface. The design focused on establishing shared vocabulary and entry points before addressing individual product UX.

Methods

  1. 1.

    Cross-product workflow mapping

  2. 2.

    User interviews across all three product user bases

  3. 3.

    Stakeholder alignment sessions with three separate product teams

  4. 4.

    Competitive analysis of automation platform consolidations

Users who learned one automation product first rarely explored the others — not because they didn't need them, but because the entry points and terminology didn't connect. The consolidation needed to solve the discovery problem before the usability problem.

User Interviews

Three separate canvas-based building experiences created a pattern proliferation problem. This became the seed of a longer-term design strategy revisited during Agent Builder.

Design Retrospective

How we worked

Process

1

Cross-Product Discovery · 3 weeks

Mapped the UX patterns, mental models, and user workflows across all three products to identify where friction occurred at the seams.

2

Unified IA & Entry Points · 5 weeks

Designed a shared information architecture and entry point system that let users navigate between products without losing context — the connecting tissue strategy.

3

Team Coordination & Delivery · 8 weeks

Coordinated design work across an 8-person team in multiple geographies, scoping and prioritizing work items while managing three separate product team stakeholders.

4

Launch & Iteration · Ongoing

Shipped the unified experience inside watsonx Orchestrate. Observed early adoption patterns to inform the Agent Builder strategy that would follow.

watsonx Orchestrate — Nov 2023 Release

watsonx Orchestrate — Nov 2023 Release

Impact

Outcomes

Summary

The Unified Automation Builder shipped inside watsonx Orchestrate, consolidating three separate products into one coherent experience. The HR promotion workflow automation alone saved 12,000 hours per year, and consistent, traceable execution reduced compliance risks across enterprise customers.

12K hrs/year

HR workflow automation — annual time savings

−10% errors

Manual error rate

You and I have had a great blend of tactical, UI-focused discussions regarding our users, their needs, and the various implementation challenges we've seen, as well as broad and strategic conversations as to where to take some of the solutions you have influenced and lead.

Design Manager, watsonx Orchestrate