Grant writing took 6-8 hours. VA costs more than the process delivered. Knowledge trapped in the business owner's head.
A solo nonprofit consultant spent 6-8 hours per grant blueprint. Work-Smart extracted her Voice DNA, built 7 AI skills from her SOPs, configured a shared workspace. Grant blueprints dropped to 1-2 hours. Same team, same hours, 3-4x output. No retainer required.
$9M in grants secured. 6-8 hours per blueprint. The math didn't work.
Mapi Velázquez runs Joy of Impact, a nonprofit consulting practice in Miami focused on grant writing, strategic planning, and capacity-building workshops for mission-driven organizations. She has secured over $9M in grants across 20+ organizations. The problem was never the quality of her work, it was how long everything took.
Her flagship service, the Grant Application Blueprint, required 6-8 hours per engagement. Eight steps: funder research, needs assessment, logic model, outcomes framework, data management plan, budget framework, narrative structure, final documentation. Strategic planning post-work after a workshop took 2-3 days. The work was repeatable in structure but custom in content, which made it feel like it couldn't be automated.
Her team was lean: Ivone, a VA in Colombia handling operations, admin, and project setup. Ale for research and data entry. Catalina designing final deliverables in Canva. But at $25-30 per hour for blueprint work, the VA cost more than the value the process delivered. The math didn't work: pay someone $150-240 per blueprint when the blueprint itself was part of a larger engagement that needed to stay affordable for nonprofit clients. Raising prices meant losing the clients she wanted to serve. Not raising prices meant burning out.
Mapi had started using AI on her own, she had 6 separate ChatGPT GPTs for different tasks. But each one lost context between conversations. No centralized knowledge base. No way for Ivone to access the same prompts and reference material. She was getting 30% of the value AI could deliver because nobody had shown her how to set it up as a system.
Deep expertise. Repeatable processes. Nothing structured for AI to help with.
We reviewed every file in her Google Drive, every proposal, every workshop deck, every email template, every strategic plan. The diagnostic revealed a common pattern for solo consultants: the expertise was deep, the processes were repeatable, but nothing was structured for AI to help with.
Knowledge Base Layer
Grant research, funder databases, proposal templates, workshop frameworks, 17 email templates, discovery call scripts, all existed but scattered across Google Drive folders, email threads, and Mapi's memory. No centralized knowledge base. No structured reference material the AI could access.
Workflow Automation Layer
Every grant blueprint followed the same 8-step structure. Every strategic plan followed the same post-workshop arc. Every workshop had the same facilitation planning sequence. These were automatable workflows disguised as custom work. Six separate ChatGPT GPTs doing fragments of what one properly configured system could handle end-to-end.
Governance Layer
Mapi was pasting client information into AI tools with no policy, no access controls, and no documentation for Ivone on what was allowed. Solo practitioners adopt AI tools fast but skip this critical layer entirely.
The fix was not a major build. It was a deep extraction and configuration: pull the expertise out of her files, structure it for AI, and set up a system both she and her VA could run independently.
Mapi Velazquez, founder of Joy of Impact
Voice DNA Extraction
We read every proposal, strategic plan, workshop deck, email, and social media post Mapi has ever written. From that, we extracted her Voice DNA, a structural communication profile that captures her sentence patterns, her sign-offs, her favorite phrases, her tone shifts between grant writing and workshop facilitation. Every AI output now sounds like Mapi wrote it, not like a template generated it. The voice profile is 280 lines of documented patterns.
15 Reference Documents (Knowledge Base)
Turned Mapi's scattered expertise into structured reference files the AI can access: a complete Grant Application Blueprint SOP (8 steps), Strategic Planning SOP, Workshop Design SOP, Client Onboarding SOP, Social Media Content SOP, Grant Research SOP, 17 email templates across 8 categories, a discovery call guide, two real completed GABs as quality benchmarks, a full services overview with team roles and tool stack, newsletter examples with voice analysis, and brand design tokens. Total: 118KB of institutional knowledge that used to live in Mapi's head and scattered Drive folders.
7 Pre-Built AI Skills
Built seven ready-to-run workflows covering every repeatable process in her practice: Meeting Debrief (paste a transcript, get summary + tasks + follow-up email in Mapi's voice), Email Composer (situation in, draft out), GAB First Draft (client brief in, 10-section grant blueprint out, the 6-8 hour process in under 2 hours), Newsletter Composer, Discovery Call to Next Steps, Proposal Generator, and Workshop Creator (covering the full workshop lifecycle from proposal through delivery prep to post-workshop reporting).
Shared Workspace Architecture
Configured a Claude workspace synced via Google Drive between Mapi and Ivone (her VA). Both access the same instructions, the same Voice DNA, the same 15 reference documents, but run independent conversations. When Mapi updates a process, Ivone sees it immediately. When Ivone runs a grant blueprint, the output matches Mapi's standards because the voice and methodology are baked into the system.
AI Governance
Established clear policies: what client data enters the AI workspace, what stays out, what Ivone can handle independently, what requires Mapi's judgment. Sensitive information, donor relationships, board dynamics, internal organizational challenges, stays human. The system handles structure and research. Strategy stays with Mapi.
Hands-On Workshop
Delivered a hands-on Cowork workshop for Mapi and Ivone. 20% presentation, 80% hands-on inside the system. Both tested real shortcuts with real transcripts, real scenarios, real email situations. The credibility moment: when the newsletter and GAB skills ran on real content and the output matched Mapi's voice exactly. Both were using the system independently before the session ended, and they scheduled a practice session for the next day on their own initiative.
Grant blueprints
6-8 hrs
→1-2 hrs
Meeting debriefs
Hours
→Minutes
VA output
1x
→3-4x
Setup time
,
→2-3 weeks
- ▸Mapi's practice went from a model where growth meant either raising prices or burning out to a model where the same team handles more clients at the same quality level.
- ▸Ivone runs the AI workflows independently, the system produces output that matches Mapi's voice because the Voice DNA and methodology are baked in. Mapi reviews and signs off. The bottleneck moved from production to quality control, which is where a business owner's time should go.
- ▸The 6 ChatGPT GPTs that used to lose context between conversations are replaced by one system with persistent memory and 7 pre-built skills.
- ▸The knowledge base that lived in Mapi's head is now searchable, shareable, and usable by every team member.
- ▸Self-contained. No monthly retainer required. Mapi and Ivone operate the workspace independently via shared Google Drive. All 7 skills are live and in daily use.
"Increíble, no hubiéramos podido hacer esto sin ti."
Questions About This Case Study
2-3 weeks from kickoff to a fully operational workspace. Mapi's VA was using the new system the same week the setup was completed. No ramp-up period, the documentation was designed so the VA could execute immediately.
This was an AI Ops Setup. Work-Smart's entry-level engagement designed specifically for solo consultants and small practices. The investment covers Voice DNA extraction, 15 reference documents structuring your knowledge base, 7 pre-built AI skills covering every repeatable workflow, shared workspace architecture, and hands-on workshop delivery. No monthly retainer required, the system runs independently once configured.
No. The AI Ops Setup exists specifically for this. Solo consultants and small practices have the same structural problem as larger companies, repeatable workflows that eat hours every week, they just need a lighter solution. If your practice has 3-5 workflows that follow a consistent pattern, this is designed for you.
Yes. Grant writing, strategic planning, coaching, financial advisory, legal intake, any practice with repeatable deliverables and a knowledge base benefits from the same approach. The specific prompts and tool connections change. The structure is the same.
No. The workspace is designed for non-technical users. The VA documentation includes step-by-step instructions. If you can write an email, you can use the system. The AI training workshop adds deeper skills if you want them, but it is not required.
The AI governance layer addresses this directly. We established clear policies on what data enters the AI workspace and what stays out. Sensitive client information, donor relationships, board dynamics, internal organizational challenges, stays in Mapi's judgment, not in the AI. The system handles structure and research. Strategy stays human.
Because the knowledge already existed, in Mapi's head, in her proposals, in her workshop decks. The setup extracted that knowledge into a communication profile (280 lines from all her documents) and built 7 pre-configured AI skills: meeting debriefs, email composer, grant blueprint drafts, newsletter composer, proposal generator, workshop creator, and discovery call follow-ups. The investment was small because the input was rich.
Yes. Mapi's VA now operates the system independently. The workspace syncs with Google Drive. Each AI skill has a clear input-to-output flow. You don't need to write prompts or understand AI. You describe what you need in plain language and the system produces a draft in your voice. The workshop was 20% presentation, 80% hands-on practice.
Spending more time on process than on the work that actually matters?
That is exactly where Joy of Impact was. Solo consultants and small practices lose hours every week to repeatable work that AI can handle, if someone sets it up properly. The AI Ops Setup is 2-3 weeks of work, and you will know the difference the first week.