Selvedge
Get started

Client: Cloud accounting firm

Building an AI-fluent accounting firm

The problem

The company’s leaders are AI enthusiasts and understood the importance of spreading that knowledge across their fully distributed team of 60 accounting professionals. The question was one of organizational structure: how to architect access to AI and upskilling in a way that’s consistent with the economic model of a professional services firm?

The solution and impact

AI knowledge in the hands of a fewEvery team member an AI champion

Treasury AP and AR Bookkeeping Payroll Client advisory Tax Administration Customer success (1 elected AI champion per division) (1 elected AI champion per division) (1 elected AI champion per division) (1 elected AI champion per division) (1 elected AI champion per division) (1 elected AI champion per division) (1 elected AI champion per division) (1 elected AI champion per division) Head office decides which AI tools are worth building Special projects builds every tool Every idea in a single queue (each team member is an AI champion) (each team member is an AI champion) (each team member is an AI champion) (each team member is an AI champion) (each team member is an AI champion) (each team member is an AI champion) (each team member is an AI champion) (each team member is an AI champion) CLIENT-FACING TOOLS Built by the team itself Billed to the client SPECIAL PROJECTS Builds internal tools used across the firm Company-wide ideas only: head office assesses
  • A team member
  • An AI champion
  • A bottleneck
  • 8 AI champions, out of 60 people
  • 1 route for every idea: through head office
  • Every tool built by the special projects team
  • 60 AI champions: every member of the team
  • 2 routes for an idea: built by the team, or by special projects
  • Only company-wide tools go to special projects

  1. Listen

    We interviewed team members to understand where the greatest opportunities for AI lay. Most of the suggested builds served individual clients, not the entire firm.

  2. Reframe

    Concentrating AI knowledge in a few hands hurts morale and long-term health. The goal: stay relevant and profitable while remaining a great place to work.

  3. Equip

    Everyone received the same AI tool and a playbook for scoping builds. Client-facing tools are built by the team and billed to the client. Only company-wide tools go to head office, then to special projects.

Read the diagram as text

Before: AI knowledge in the hands of a few

  1. Each of eight divisions elects a single AI champion.
  2. Champions send every idea to head office, which decides which AI tools are worth building.
  3. The special projects team builds every tool, so every idea waits in a single queue.

After: Every team member an AI champion

  1. Every team member in every division is an AI champion.
  2. Client-facing tools are built by the team itself and billed to the client.
  3. Only company-wide ideas go to head office, which assesses them.
  4. The special projects team builds internal tools used across the firm.

Every engagement begins with a conversation.

Get started