Talent infrastructure for an AI-shaped workforce.
We run AI agents across our own operation, so we hire for the operating model our clients are moving into, not the one they're leaving.
PathwaysX is a team of operators, researchers and evaluators. Every search is led by a human Talent Partner who is also an AI practitioner, with accountable human judgement behind every finalist.
Pathways Connect, our own AI product, structures sourcing, assessment and evidence. We continuously map the talent markets we cover, then activate the relevant network for each search rather than starting from zero.
- Finalists per search
- Three to five, in two to three weeks.
- Evidence standard
- A structured Decision Pack on every finalist.
- Search leadership
- A human Talent Partner, end to end.
- Operating system
- Pathways Connect structures the work; judgement stays human.
The work is being rebuilt for agentic operating models. Hiring hasn't caught up.
Across the market, teams are moving AI agents from pilots into production. The roles that carry that work barely existed two years ago, and the hiring playbook for them is still being written. That's the gap we work in.
- 2025 → 2026
Agents move from pilot to production.
Enterprises stop demoing agents and start rebuilding workflows around them. New operating roles appear between the AI capability and the business function it serves.
- The postings already carry it
AI fluency shows up where it never did before.
Job postings for operations, marketing and finance roles now ask for AI fluency, language that wasn't there 18 months ago. The evaluation methods haven't kept pace with the postings.
- New titles, old functions
Hybrid roles resist conventional screening.
Forward-deployed engineers, AI transformation leads, agent operations managers: titles that combine two disciplines and defeat keyword screening.
Where the new roles sit
Swipe to see all five columns →
Each of these maps to one of the five families we hire against.
Three hiring consequences follow from that shift.
The work is rebuilt around agentic ops so the roles are too.
Teams aren't adding an AI person next to the existing org chart. They're re-cutting responsibilities around what agents now do. That means the job descriptions everyone is hiring against are out of date.
Density sits between functions not inside them.
The scarce profiles combine two disciplines: product judgement with model fluency, platform depth with operational range. Screening for one keyword finds neither.
The pool is narrow by construction combinations, not credentials.
There is no degree for 'runs an agentic operating model'. The people who can do this work are found by how they're defined and evaluated, not by how they're filtered.
Every search runs the same four stages. Nothing starts until the first one is done.
Most hiring failure is designed in before the first candidate call: a vague description, an unweighted rubric, an interviewer making it up as they go. We fix the design first, then run the search against it.
- 01
Outcome
We start with the business result the role must deliver in 12 months, agreed with every stakeholder before anything is written.
- 02
Capability
The outcome becomes a weighted capability profile: the skills, judgement and range the hire actually needs, scored before sourcing begins.
- 03
Talent Model
We map where that profile sits in the market: which functions, which companies, and what it takes to move those people.
- 04
Role
Only then is the role written: calibrated to the market, priced correctly, and ready to run a structured search against.
Five families we hire against: the roles AI is rewriting across the functions we know best.
We hire two kinds of roles: people who build AI, and people in the business whose work is now run with AI.
Across Technology, Operations, Marketing, Partnerships, Customer Success and Finance, the most important roles often cross two families.
| Family | Representative roles |
|---|---|
| Applied AI & product engineering | Founding AI Product Lead · AI Integration Architect · Forward-Deployed Engineer · ML / Applied AI Engineer |
| AI product & delivery leadership | VP / Head of AI Product · Director, AI Transformation · Group Product Manager, Applied AI |
| Platform & infrastructure | Head of Developer Platform · Agent-runtime / Platform leads · MLOps & Evaluation leads |
| Data, governance & security | Director, AI Data & Analytics · AI Governance Lead · Data platform & analytics leads |
| Business AI roles | Revenue Operations Lead · Marketing Operations Lead · Customer Success AI Lead · Finance AI Lead · AI Enablement Lead |
Three to five finalists in two to three weeks, each with a Decision Pack you can interrogate.
A Decision Pack is a structured evidence file on every finalist: how they scored against the profile, where the evidence came from, what the trade-offs are. It turns a hiring debate into a hiring decision.
A human Talent Partner leads every search and presents the pack with you. No black-box shortlists.
While a search is open, you receive a report every two weeks. Each quarter, we can deliver a tailored market report on how AI is changing the work, roles and pay in your industry.
Every search closes with the same question answered: of the people we could have hired, why this one, and what did we give up?
Transparent, risk-aligned pricing. Fees tied to milestones or 12-month retention.
No locked-in retainers and no opaque percentages. Every model is published, and you can test the business case against a traditional agency before we speak.
Start a conversation
If you're hiring where the role doesn't have a name yet, we should talk.
Need org-wide AI fluency, not only a hire? PathwaysX runs workshops with your team, and can assemble an implementation team when something needs to ship.