Director, AI Orchestration
About the role
The Director, AI Orchestration owns the system that makes Togetherwork AI-native. The core of the job is replacing manual work with agentic workflows across the company, so people spend their time on the judgment, relationships, and decisions where humans matter most.
Responsibilities
Own the vision and ongoing curation of Togetherwork's enterprise knowledge infrastructure (in coordination with the CTO and security team): the system that brings together product signals, support data, customer trends, and operational knowledge so agents and humans across every function can draw on it reliably.
Define standards for how each function contributes to and maintains its knowledge domain, so the infrastructure stays current and accurate as the business evolves.
Measure knowledge infrastructure health: coverage, freshness, retrieval quality, and actual usage by agents across the company.
Set the standards every team builds agents to: design patterns, quality bar, human-in-the-loop checkpoints, testing, deployment, and retirement.
Select and run the shared agentic tooling the company builds on, staying tool agnostic where possible and documenting the tradeoff whenever lock-in is the right call.
Keep standards current as tooling and model capabilities change, and retire patterns that stop earning their place.
Define how agent quality is measured beyond business outcomes: reliability, accuracy, latency, and failure modes, so the company knows when an agent is working and when it isn't.
Identify the manual workflows across the company where agents can take over the repetitive work, prioritized by cost saved or revenue generated fastest.
Test each workflow for the level of orchestration it can handle, from AI-assisted to fully agentic, and be honest about which is which.
Build the first proof cases hands-on, then harden winning prototypes into production workflows with clear owners, human checkpoints, and escalation paths.
Track every replaced workflow: what the agents took over, what the humans moved on to, and what it returned.
Help organize the AI guild: Builders assigned by each ELT member from their function, working on shared standards and tooling.
Run hackathons that map cross-functional workflows piece by piece and test levels of agent orchestration against real work.
Turn hackathon wins into production workflows and hackathon lessons into updated standards.
Train and certify guild Builders so each function can build and run its own agentic workflows; step in hands-on when a function stalls, then hand it back.
Respect functional ownership: each ELT member decides what is best for their function and remains accountable for the quality and stewardship of their function's data and content; you provide the standards, training, and guardrails that keep it coherent company-wide.
Own access controls at the policy level: which agents reach which data and systems, with least-privilege defaults and clear separation between product and enterprise data.
Define the rules for agent autonomy: what agents may do unattended, what requires human sign-off, and how that boundary moves as trust is earned.
Watch for excess risk across the AI portfolio, including data exposure, vendor concentration, compliance, and customer impact, and raise it early with a plan.
Own the communication rhythm for the AI program: what is changing, why, what it means for each team, and where to get help, in partnership with Communications.
Assess organizational readiness before each workflow goes live: the affected team knows what is changing, has been trained, and knows where the human checkpoints sit.
Manage the AI Program Manager, who runs the AI PMO: a single, prioritized view of AI initiatives across the company with owners, timelines, and expected impact.
Define and report the scoreboard: workflows replaced, hours of manual work eliminated, adoption by function, governance compliance, cost saved, revenue influenced, and cycle time reduced.
Hold functions accountable to committed outcomes and report progress and risks to the AI Council Lead and ELT in plain terms with the data behind each call.
Kill or redirect initiatives that are not delivering, quickly and without drama.
Track and manage the cost of AI operations: model usage, token spend, tooling licenses, and infrastructure, with a bias toward doing more with less as the portfolio scales.
Qualifications
Hands-on experience replacing real business processes with agentic or automated workflows, and the judgment to know which processes are ready and which are not. This includes real work application and use of low-code/no-code tools and AI platforms (Claude, GPT, LangGraph, n8n, and others).
7+ years in product, program, or operations leadership, including direct people management with a very strong focus on change management.
Experience running a hub-and-spoke, guild, or center-of-excellence model: setting standards centrally while enabling distributed teams to execute.
Deep appreciation for cross-functional business processes and how work gets done in functions like HR, Legal, Finance, Support, Professional & Managed Services, Revenue, and CX.
Demonstrated ability to influence executives and work across functions without direct authority, letting each leader own their domain.
Experience with AI governance: usage policy, tool evaluation, access decisions, agent autonomy boundaries, and risk review.
Tool-agnostic mindset: able to evaluate platforms on merit, avoid unnecessary lock-in, and explain the tradeoffs in business terms.
Track record measuring and reporting program outcomes, and the willingness to shut down work that is not delivering.
Comfortable as a working leader: managing a team while still building workflows, standards, and proof cases directly.
Must be in Austin, TX. Hybrid work requires 2-3 days in office.