AI Readiness Assessment · Anonymous Sample
Process mapping · Bottleneck identification · Stack evaluation · Operational action plan
Section 01 · Executive Summary
Acme Recruitment is a fast-growing executive search agency (+38% revenue YoY) with a structural problem: 56% of recruiter team time goes to repetitive administrative tasks — CV screening, data entry on legacy ATS, interview scheduling, manual client reporting. Management knows they need freed-up hours to scale commercially, but internal initiatives (ChatGPT prompt templates, sporadic Make automations) have not generated systemic impact.
Recommendation: proceed with a Co-Building Sprint of 5-7 weeks focused on the 3 priority processes (see Section 04). Estimated investment €34,000-€42,000, expected payback 7-10 weeks, "hours recovered or refund" guarantee included.
Section 02 · Operational Context
The team uses a fragmented stack: Bullhorn as primary ATS, Microsoft 365 for email/calendar, Slack internally, scattered Google Sheets for commercial tracking, Make for 2-3 isolated automations. No intelligence layer on top of the ATS.
Recruiter downloads CVs from Bullhorn, reads manually, evaluates fit against JD, rejects or advances. 70% rejected after 3-5 min of reading. No fit-scoring automation.
Manual email personalization on LinkedIn + Bullhorn. Reply rate 6%. Generic templates hurt the brand.
Back-and-forth email coordination between candidate, client, and internal team. Calendly barely used because enterprise clients don't want public booking links.
Manual Excel files updated every Friday. Candidate status, pipeline, next steps. Duplicate of data already in Bullhorn.
PDF/DOC received via email. Hard/soft requirement extraction, reformatting for Bullhorn. Frequent errors.
Identification of companies in hiring spike. Mixed public sources, manual, not scalable.
Section 03 · Prioritized Bottlenecks
We ranked 6 bottlenecks by time impact, technical feasibility, and speed to steady state. Only the top 3 warrant an immediate sprint.
| Bottleneck | Impact (h/week) | Feasibility | Priority |
|---|---|---|---|
| Inbound CV screening Agent: JD-driven fit-scoring + classification | 168 h/week | High. Bullhorn has API | P0 |
| Personalized outreach Agent: drafting + multi-touch sequence | 108 h/week | High, LinkedIn Sales Nav integration | P0 |
| Weekly reporting Agent: Bullhorn extraction → narrative client report | 48 h/week | High, direct on Bullhorn | P0 |
| Interview scheduling Agent: email-based coordinator | 60 h/week | Medium, requires 3-way orchestration | P1 |
| JD parsing Agent: intake doc + Bullhorn fill | 36 h/week | High, low volume | P1 |
| BD prospecting Agent: companies-in-hiring spotter | 24 h/week | Low, depends on unstable external sources | P2 |
Focusing the sprint on the 3 P0s recovers 324 h/week in total (~57% of the team's administrative time). We tackle P1s in a subsequent sprint, once the 3 P0 agents are running stably.
Section 04 · Stack Assessment
| Tool | Status | Recommendation |
|---|---|---|
| Bullhorn (ATS) | Keep | Remains source of truth. Agents write via API. |
| Microsoft 365 | Keep | Email + calendar OK. Add Graph API for scheduling agent. |
| Make (automations) | Reposition | Stays for light glue. Complex agents move to a dedicated layer. |
| Google Sheets tracking | Remove | Duplicate of Bullhorn. Replaced by automated reporting. |
| LinkedIn Sales Navigator | Keep | Required for outreach agent + BD prospecting. |
| AI agents layer (new) | Add | Company LLM provider (Anthropic Claude Business) + workflow orchestrator + observability. Stack agnostic, details in Action Plan. |
Compliance: candidates are personal data. The technical proposal includes EU infrastructure (Cloudflare Workers + EU-hosted LLM), full audit log of every agent decision, configurable retention (default 90 days post-closure), Art. 28 DPA with Soraia included.
Section 05 · Operational Action Plan
Full-day session with sponsor (CEO + COO + lead recruiter). Lock targets for the 3 P0 agents with measurable metrics. Dev environment setup + Bullhorn API access.
Build fit-scoring agent against JD. Test on historical dataset from the last 6 months. Threshold calibration with 3 senior recruiters. Deploy in shadow mode (parallel to manual screening).
Email drafting personalization agent. Brand-approved templates. Multi-touch sequence (3 steps). LinkedIn Sales Nav + Bullhorn outbound integration.
Pipeline extraction from Bullhorn → client-facing narrative report. Custom template per client. Auto-send Friday 5pm with human preview in approval queue.
4-hour workshop with all 18 team members. Agent demos, governance, escalation rules. Definition of internal "agent champion" (1 senior recruiter).
Shadow mode switched off. Week-1 measurement vs baseline. Final agent adjustments. 30 days of hypercare with weekly check-ins.
Section 06 · Investment + Guarantee
| Item | Weeks | Amount |
|---|---|---|
| Co-Building Sprint (3 P0 agents) | 5 | €28,000-€34,000 |
| AI Adoption workshop (18 people, 1 batch) | 1 | €5,000 |
| Hypercare (30 days post go-live) | 4 | Included |
| Total | 5-7 weeks | €33,000-€39,000 |
"Hours recovered or refund" guarantee. Sprint target: average recovery of 7 hours/week per recruiter at go-live + 30 days of hypercare. If we don't hit it, we work for free until we do, or we refund the sprint. Measurement against the baseline defined in Section 02.
Next Steps
To align on next steps: [email protected] · [email protected] · soraia.io/en/contact.
This document is an anonymized sample. All client data, operational metrics, and projections are indicative of a real Soraia case but have been anonymized. The actual assessment is custom for each client and includes data measured directly against their operations.