AI Consulting

AI consulting that starts with the work

We start with how your team gets work done today, find the steps worth changing, and only then decide whether AI belongs there. We are vendor-agnostic, with no partnerships, commissions, or referral arrangements with AI platform vendors.

AI on the office floor

Automation

The system does the step for you. Read more about Automation

Augmentation

Your people do the work, with AI assisting. Read more about Augmentation

Agency

An AI agent works on your behalf, under supervision. Read more about Agency

People stay in charge

Our founding assumption is that every agentic workflow is supervised by a named person. Read more about People stay in charge

Security from the first day

An AI tool that can read your documents or act inside your systems is also a new way in for an attacker. Read more about Security from the first day

Three ways a workflow can change

Automation, augmentation, and agency

Once we know which workflow to change, AI can change it in one of three ways. Most organizations end up with some of each.

Automation

The system does the step for you

Automation follows explicit instructions to do repetitive work the same way every time. It fits steps where the rules are clear and the volume is high.

Where it fits

  • Moving data between the systems you already use
  • Reports that compile and send themselves on schedule
  • Reading, sorting, and routing documents and forms

Augmentation

Your people do the work, with AI assisting

Augmentation is a collaboration between a person and an AI tool. The person keeps the judgment. The tool drafts, searches, summarizes, and checks.

Where it fits

  • Drafting and editing in your organization's voice
  • Search across your own documents and records
  • Analysis that supports a decision someone still has to make

Agency

An AI agent works on your behalf, under supervision

Agency means an AI agent carries out a multi-step task on its own. We only design these with a named person who supervises the agent, reviews its output, and can stop it.

Where it fits

  • First-line customer questions, with a handoff to a person
  • Research and summaries for a person to review
  • Routine back-office tasks with an approval step

People stay in charge

Every AI workflow has a named supervisor

Our founding assumption is that every agentic workflow is supervised by a named person. That person reviews what the workflow produces, answers for its mistakes, and can switch it off.

We also keep the load human-sized. When the work grows, we add supervisors rather than unsupervised agents.

Security from the first day

An AI tool that can read your documents or act inside your systems is also a new way in for an attacker. Security is part of every AI engagement, not a phase at the end.

Security assessment

We review how your data, systems, and AI tools are exposed today and rank what to fix first.

Incident response planning

A written plan for who does what when something goes wrong, so your team is not deciding under pressure.

Security architecture for AI

Access, data handling, and logging designed into each AI workflow before it goes live.

New tools change how people work, and someone has to lead that change. Our management services prepare your managers and staff for it.

Management and Change Leadership

Find the workflow worth changing first

Tell us where your team loses time, or where AI tools are already showing up unannounced. We will tell you what is worth changing and what is not.