Eldris

Research and judgment for putting AI to work.

Eldris is an applied AI firm. We advise leaders on consequential AI decisions, study how people actually come to use AI, and build OOMU, software that puts it to work on the Mac you already own.

Why Eldris

Every side of the table.

Eldris is led by Dr. Jeffrey Allan, who built OOMU from the ground up and directs the Institute for Responsible Technology and Artificial Intelligence at Nazareth University. Behind that sit fifteen years working in Asia, senior roles in multinational companies, two technology startup exits, and years of state and federal technology policy work.

A question about AI rarely stays technical, commercial, or political for long. The firm is built to follow it across all three.

What we do

Three ways to work with Eldris.

Each one starts from a different situation and ends with something you can use.

Advisory

A decision that has to hold up.

For leaders, boards, and the advisors around them. You receive an independent assessment, a decision memo or executive briefing, and the assumptions behind it stated plainly.

Research

A question that needs evidence.

Original studies of how people and organizations come to use AI, commissioned for your question or licensed for internal use.

OOMU®

AI work that runs on your Mac.

A native Mac app, free to use, including at work. For organizations, OOMU Control governs it centrally, and we can help with setup, workflow design, and training.

Research

The gap between having AI and using it.

Eldris research is organized in two programs. The Enterprise AI Adoption Gap looks at people inside organizations, where roles, incentives, accountability, and coordination shape what gets used. The Consumer AI Adoption Gap looks at individuals, whose circumstances and tasks differ from one person to the next.

Each program is a series of bounded studies, and each study defines a decision, a population, a task, and a boundary before any fieldwork begins.

Illustrative question, not a finalized study

What keeps an employee who has access to AI from using it for a particular work task?

Advisory

Bring us a consequential AI decision.

  • Deciding where AI is worth using A scoped assessment of where it pays, where it does not, and what it would take.
  • Testing a proposal or a vendor An independent review of the claims, the costs, and the risks, delivered as a decision memo.
  • Preparing leaders or a board An executive briefing that leaves the room able to ask the right questions and weigh the answers.
  • Overseeing delivery Senior oversight across technical teams and markets, so the work stays tied to the decision it was meant to serve.

Senior-led and team-backed. Our assessment follows the evidence, including when it does not support the preferred direction.

OOMU in use

AI that works on the Mac you already own.

You ask in your own words. For each request, OOMU decides on your Mac whether the Mac or a cloud model should answer, within the same chat, without you switching models.

The OOMU window, asking "What would you like to do?" with three starting jobs, Weekly Brief, Summarize Folder, and Help with Email, and a box that says "Ask OOMU anything." Below it, a switch set to Auto, with On this Mac and Cloud as the other choices.
OOMU Beta 3. The switch under the message box sets the frame: Auto, On this Mac, or Cloud. Cloud models stay out of Auto until you turn them on, and when you use one, your provider bills you for it.

The app

Free, including at work.

The whole app, for commercial use too. Everyday work runs on your Mac with no charge per question. Optional cloud models are billed by the provider you choose.

OOMU Control

Paid, for organizations.

The server an organization runs to deploy OOMU, decide which models every Mac may use, and see what each one is running and what it blocked.

Implementation

Paid, scoped to the work.

Setup, workflow design, and training for a team putting OOMU to work, agreed in writing before it starts.

What are you working on?

Pick the closest route. A short description is enough to start, and we can define the scope together.