Business Analyst, AI Workflows

ECI Software Solutions•Westlake, TX

15h•Remote

About The Position

Global IT's Project Management Office runs ECI's strategic delivery portfolio – M&A integrations, platform consolidations, compliance programs, and the cross-cutting initiatives that ride across every function. AI-enabled efficiency is one of the most important arcs of work on that portfolio, and we are investing in dedicated execution capacity to match the opportunity. We believe AI transformation is built, not bought. The work that creates lasting value isn't the tool selection – it's mapping the workflow end-to-end, capturing the tribal knowledge that lives in people's heads, and then designing, building, and running the solution alongside the teams who own the work. That requires a different shape of role than the traditional separation of Business Analyst, developer, and Project Manager, because the handoffs between three specialists are where momentum and context tend to be lost. The Business Analyst, AI Workflows is that integrator. The role sits inside IT PMO, partners across Platform, Engineering, Applications, Support, HR, and Finance, and is accountable for translating ECI's AI strategy into a real portfolio of measurable, shipped automations that serve our customers and our teams.

Requirements

5+ years combined experience spanning business analysis, automation or low-code implementation, and project or program management. Expertise in at least two of these areas is a must.
Demonstrated ability to map a cross-functional workflow end-to-end, including the exceptions and undocumented decision points, and produce a specification an engineer or an agent can act on.
Hands-on experience building automations in at least two of: Salesforce Flow, Power Automate, n8n, Zapier-class platforms, Claude Code, or comparable low-code tooling.
Direct experience shipping an automation or AI workflow to production with a measured before-and-after comparison – not a pilot that lived only in UAT.
Disciplined project cadence: writes and runs status updates, holds phase gates, measures outcomes against committed numbers, and communicates risks early.
Strong communication skills with the ability to explain a workflow, a measurement, and a tradeoff to both an engineering audience and an executive audience.
Comfort working across SaaS systems and integrating them via APIs; able to read documentation and prototype an integration without waiting for engineering support.

Nice To Haves

Experience with Salesforce administration and Salesforce Flow at scale.
Hands-on work with agentic AI tooling (LLM-backed assistants, Salesforce Agentforce, internal copilots, MCP-style integrations) – including the operational reality of guardrails, evaluation, and failure modes.
SQL fluency; light Python or JavaScript scripting; familiarity with REST APIs and webhooks.
Familiarity with ITIL service-management practices or PMP-style project disciplines.
Prior experience inside a serial-acquirer or multi-business-unit software company, where workflows differ meaningfully between business units.

Responsibilities

Map current-state workflows end-to-end across IT, Support, Finance, HR, and Engineering – including the exceptions and edge cases that live in operators' heads, not in the documentation.
Conduct working sessions with process owners and frontline staff to capture tribal knowledge and convert it into rules, decision logic, and instructions that automation or agents can follow.
Gather requirements and write clear specifications that include success metrics, failure modes, and the exact human-in-the-loop checkpoints required for safe rollout.
Apply a triage methodology to every candidate workflow – deterministic automation, agent-handled judgment, or humans-only – and document the rationale so triage decisions are auditable.
Establish and maintain the baseline measurement behind ECI's AI workflow commitments so progress is reported on real numbers, not estimates.
Build the candidate-workflow portfolio against ECI's evaluation criteria: frequency, repeatability of decisions, context spread across systems, and measurable pain.
Implement automations directly in Salesforce Flow, low-code and no-code platforms (Power Automate, Zapier-class tools, internal Abacus capabilities), and emerging agentic AI tooling.
Write light scripting (Python, JavaScript, SQL) and integrate across SaaS systems via REST APIs and webhooks where no-code tooling does not reach.
Configure agent guardrails, evaluation prompts, and feedback loops; design for graceful failure when the agent is uncertain rather than for confident wrong answers.
Partner with Platform Engineering and Application teams to ship integrations safely on top of existing systems (Salesforce, Vonage, Workday, JSM) – without forcing migrations.
Build the instrumentation that proves an automation is working – logging, dashboards, and drift detection – so shadow-mode and production performance can be compared on the same axes.
Maintain a clean, documented build inventory so any automation can be handed off, audited, or rolled back without tribal knowledge becoming the dependency we just removed somewhere else.
Run each workflow through a phased delivery model: sandbox build, shadow mode (agent observes, humans decide), supervised production (agent acts, humans review), steady state.
Enforce human-in-the-loop discipline at every phase boundary; no automation moves to the next stage without measured evidence and an owner sign-off.
Own the regular progress reporting on AI workflow delivery, including which workflows shipped, the measured impact of each, and a credible forecast for the next quarter.
Coordinate with other technical resources when an AI workflow intersects an active Initiative – M&A integration playbooks, platform consolidations, compliance projects.
Partner with process owners across functions to land each automation with the people who do the work – they are the success criteria, not a downstream consumer.
Maintain a single source of truth for the workflow portfolio (status, phase, baseline, current measurement, next checkpoint) and surface it on a recurring cadence to the IT leadership group.