For consultancies

Polaris for consultancies

The scarce resource in a consultancy is not hours. It is partner attention, spent on work a partner should not be doing.

The short answer

Consultancies sell judgement, but every engagement begins with a research and synthesis floor that consumes senior time before any judgement is possible. Polaris moves that layer to AI workers that run on cloud machines, use live web search, and deliver the output as files attached to a task. Hours are estimated by an open formula and logged, so the cost of the floor becomes a visible line.

Deliverables: Arrive as files on the task
Hours: Open formula, itemised per job
Cost per consultant seat: $0

The fixed tax on every engagement

A consultancy prices its work on expertise, then spends the first two weeks of every engagement on something that is not expertise. Reading the client's market. Assembling the comparable set. Pulling the regulatory background. Turning nineteen interviews into four themes. Building the same category map that was built for a similar client eighteen months ago and cannot be found now.

This floor is unavoidable, which is exactly why it deserves attention. It is the largest block of hours in the engagement that does not depend on who is doing it, and in most firms it is done by the people whose time is either the most expensive or the most developmentally valuable. Neither is a good outcome.

The four costs hiding inside the floor

Senior time spent reading, not deciding
The partner or principal reads to form a view. Half of that reading is orientation any competent researcher could have compressed into a briefing note.
Work already done, done again
Firms rebuild the same market maps because the previous version lives in a deck on someone's laptop rather than in a place with a search box.
Synthesis that arrives too late to change the answer
Interview themes surface in week four. The hypothesis was set in week one and the evidence is now decoration.
The proposal that costs a week to lose
Every unsuccessful pitch consumed real research hours that nobody billed and nobody logged.

Running a diagnostic engagement in Polaris

The structure is ordinary. The difference is who does the floor.

1
The engagement becomes a workstream
One bucket per engagement, holding tasks, docs, files and the conversation. Docs are a nested tree with a block editor, versioned files, review and comments, so the interim work products live where the work lives.
2
Hire a research analyst briefed on this client
The Chief of Staff runs a short interview and the answers become a readable SKILL.md. That file is the difference between a generic researcher and one that knows this client's sector, vocabulary and evidence standard.
3
Assign the floor as discrete tasks
Comparable set, regulatory background, category map, competitor positions. Each is a task with acceptance criteria that the worker ticks as it goes, so a half-finished piece of research is visibly half-finished.
4
Deliverables arrive as files
The machine posts its output as a comment on the task, including generated files such as PDFs and documents. The consultant reads a document rather than a chat transcript.
5
The principal spends their time on the view
A human closes and rates every task. The judgement, the client relationship and the recommendation stay exactly where the fee assumes they are.

Where a principal's week goes

Today

Orientation reading that could have been a briefing note
Rebuilding an analysis the firm has done before
Chasing an interim deliverable from three people
Unbilled research on a proposal that may not land

With the floor delegated

Reading a sourced briefing note produced overnight
Reusing a skill file that carries the firm's method
Watching acceptance criteria tick on the task itself
A logged, itemised hour count against every proposal

Next, for a consultancy

The research role, the proposal work, and what an agent-produced file actually looks like.

Hire an AI research analyst

Assign the question on Tuesday afternoon and read the brief before your Wednesday call, with every claim carrying the URL it came from.

Hire an AI competitive analyst

Assign it every Monday and you get a written record of what your market did last week, with a URL behind every line of it.

Research memos waiting for you in the morning

The questions worth researching are the ones you never have a free afternoon for.

Proposals drafted from your call notes, not from a template

Every proposal is 70% the same document and 30% the reason this customer is different. The 30% is the part worth your evening.

When the deliverable is a file

A chat reply is not a deliverable if the thing you needed was a document somebody can open.

Questions people ask

+How is this different from the agencies argument?

An agency's constraint is the blended cost of a delivery hour across a team. A consultancy's constraint is senior attention, which cannot be blended, hired quickly or scaled. The agency case is about margin per hour. This one is about what the most expensive person in the building spends their week reading.

+Can a worker be trusted with primary research quality?

Treat it as a first pass with sources attached, not as a finished view. Workers run a live tool loop with real web search and tick their own acceptance criteria, and the delivery arrives as a file you read like any junior's work product. A human closes every task, so nothing reaches a client without a consultant having accepted it.

+What stops us rebuilding the same analysis for the next client?

The skill file. A worker's capabilities are a SKILL.md you can read, edit and reuse, so the method your firm applies to a market scan becomes a durable artifact rather than a habit in a partner's head. Docs are versioned and nested, so the outputs stay findable in the workstream rather than in an attachment.

+Does the hour estimate hold up under scrutiny from a finance team?

It is derived from observable effort: base pickup time, searches at roughly twelve minutes each, finished prose at about ninety characters per minute, and fixed overheads for checklist items, comments addressed and files produced, clamped between five minutes and eight hours per session. The formula is open and every job is logged, so any line can be challenged from the log itself.

+Is a small consultancy the right size for this?

Boutique firms feel the argument hardest, because the research floor lands on partners directly with no junior bench to absorb it. Larger firms have that bench, so the case becomes an economic one about what that bench costs rather than a capacity one. Both work; the boutique version is sharper.

Polaris for agencies

You already know what an hour of delivery sells for. This is a page about what it costs.