AI Executive Assistant for a Consulting Firm

The Challenge

A principal consultant at a small, specialized consulting firm was carrying the operational weight of an entire business. Competing against firms like Accenture, Deloitte, and KPMG, the quality of the work is what wins. But for a small firm to deliver at that level, one person ends up doing everything.

On any given week, that meant: prospecting and relationship nurturing across a portfolio of clients, designing and packaging new service offerings, selling those offerings to enterprise buyers, delivering multi-day training programs and executive coaching, creating polished presentations and sales decks, managing follow-ups and accountability across a distributed team, and processing a constant flow of emails, Slack messages, and calendar conflicts.

He had never had an executive assistant. Not as a consultant, not even as a VP of Product at previous companies. The ceiling was always his own time.

Key pain points:

• Every AI session started from scratch. Endless re-explaining of context, clients, and priorities.
• Output was just text in a chat window. Still required manual copy-paste, formatting, filing, and follow-up.
• No system of record. Action items, client history, and relationship threads lived in his head or scattered across tools.
• Weekend overflow was the norm. Firm-building work bled into personal time every week.

The Approach

AI didn't have to be a question-answering tool. It could be a system that retains everything about the business, the clients, and the work in progress, and operates alongside you every day without being re-briefed.

The key design decisions: build a structured knowledge vault so the AI never loses context. Integrate directly with the tools already in use. And critically, have the AI do the work, not just suggest it, with human review before anything goes out.

What we built:

• A structured knowledge vault (Obsidian) with dedicated files for every client, prospect, project, decision, and action item
• Live integrations with Microsoft Outlook (email and calendar), Microsoft To Do, and automated meeting transcription via Krisp
• Automated transcript processing: raw recordings converted to structured meeting notes with decisions, action items, and client-specific filing
• A daily command center: consolidated briefing of meetings, tasks, actionable emails, and pending items across all clients
• Branded presentation generation using established templates, formatting standards, and visual language
• Proactive accountability tracking across every client, team member, and timeline

The Results

The difference between this and a chatbot is that the system remembers. It knows the clients, the priorities, and the history. It doesn't wait to be asked. That's what eliminated the weekend overflow and cut follow-up time by 95%.

Follow-through on business development became faster and more consistent across multiple active prospects. Client delivery improved with better preparation, documentation, and continuity between engagements. A firm of seven started operating like one with a back office.

What's Next

The firm is now rolling this out to every consultant. Each one will get their own AI executive assistant, built on the same architecture, with the same persistent memory and ability to do real work.