Quick answer: Six steps stand up AI workflow automation for HR in 60 days: scope the highest-volume workflow, map the current state, pick the AI components, build the Make.com orchestration, pilot with one team, then roll out org-wide. The first 30 days produce the pilot; the second 30 days produce the rollout. Total team cost: one analyst plus 40 hours of HR business partner time across the 60 days.

Most HR AI deployments are sold by the AI vendor and built by the AI vendor’s professional services team. The result is a model that works in isolation and an HR team that does not adopt it. The fix is starting with the workflow, not the AI. This how-to walks through the six steps in the sequence that has worked across our deployments, anchored to the pillar architecture in AI Candidate Screening: A 7-Step Blueprint for Automated Hiring (2026) and the faster screening guide at Faster Candidate Screening — A Step-by-Step Implementation Guide (2026).

What do you need before you start?

Three things. One sponsor — a VP HR or CHRO who has authorized the budget and will defend the project at executive review. One workflow chosen — the highest-volume repetitive workflow in HR, usually candidate screening or onboarding. One Make.com workspace — paid tier, two seats, basic admin access.

Step 1 — Scope the workflow (week 1)

Pick the workflow with the highest annual volume and the most recruiter time spent. For most mid-market orgs this is candidate screening. Document current-state volume (requisitions per quarter, candidates per requisition), current-state time spent (recruiter hours), and current-state pain points (specific complaints from recruiters and hiring managers).

Step 2 — Map current state (week 2)

Walk through the workflow as it runs today. Where does the data come from? Where does it go? Who touches it at each step? Where are the handoffs? Where are the manual steps that consume time? Output: a visual workflow diagram and a list of automation candidates ranked by time-savings potential.

Step 3 — Pick the AI components (week 3)

For candidate screening, the AI components are a resume parser, a skills-extraction model, and a scoring model. Pick vendors using the 12-red-flag checklist. The components plug into the Make.com orchestration; the orchestration is what you own. The components are commodities and swappable. For the skills matching context, see AI Skills Matching for Recruitment — Implementation Guide (2026).

Step 4 — Build the Make.com orchestration (weeks 4-5)

Build the Make.com scenarios that move data between the ATS, the AI components, and the back to the ATS. Standard pattern: scheduled trigger on the ATS to pull new applications, route each application through the parser, route the parsed output through the scoring model, write the score back to the ATS as a structured field, log the full chain to an audit data store. Each module includes sent_from/sent_to, error handling, and named steps.

Step 5 — Pilot with one team (weeks 6-9)

Pick one recruiting team, ideally 3-5 recruiters. Run their requisitions through the new automated flow alongside the old manual flow for 30 days. Measure the seven ROI metrics. Collect feedback from recruiters and hiring managers. Fix the rough edges. Do not skip this step — every deployment that skipped it produced rework.

Step 6 — Roll out org-wide (weeks 10-12)

Once the pilot has 30 days of data showing the seven metrics moved in the right direction, roll out to the rest of the org one team per week. Weekly office hours with recruiters during rollout. By week 12, the org-wide deployment is complete and the analyst transitions from build mode to monitor mode.

What can go wrong?

Three failure modes. First, no sponsor — the deployment loses budget at the first quarterly review. Second, wrong workflow chosen — picking a low-volume workflow produces low ROI even with perfect execution. Third, vendor dependency — buying a vendor’s all-in-one platform instead of owning the Make.com orchestration leaves you locked in when the vendor lags on AI features.

What’s next

If you have the sponsor and the workflow chosen, start step 2 this week. The deliverable is a workflow diagram and an automation candidate list. Bring both to the next executive review. For the full screening architecture this sits inside, see the AI Candidate Screening: A 7-Step Blueprint for Automated Hiring (2026).

Sources

• SHRM, “HR Technology Adoption Trends,” 2025
• McKinsey, “The Future of HR Automation,” 2024
• Internal client deployment data, 2024-2025

Summary: Six steps stand up AI workflow automation for HR in 60 days — scope, map, pick components, build Make.com orchestration, pilot, roll out. Skip the workflow map at step 2 and the deployment fails.