TIER-00
Consultation
Named engineer on the call. You leave with three concrete automations, expected hours saved per week, and an honest note if AI is not the right tool.
$75
flat, 30 to 60 min
guide / pricing
AI automation consulting pricing, without the discovery call
Short version: in 2026 a paid AI automation consult runs roughly $75 to $300, a single-workflow build runs $500 to $5,000, a custom system runs $2,000 to $25,000, and a retainer runs $1,000 to $8,000 a month for small-business work. Enterprise programs run far higher. The reason you cannot simply look that up is not that the number does not exist. It is that naming it early costs the seller money.
M
Matthew Diakonov, Written with AIDirect answer / verified 2026-05-18
For small-business scale work, AI automation consulting in 2026 costs about:
- Paid consult / readiness audit$75 to $300 (free discovery calls are common too)
- Single workflow build$500 to $5,000, fixed scope
- Custom multi-step system$2,000 to $25,000, fixed scope
- Monthly retainer$1,000 to $8,000 a month, after a build
Hourly work sits at $80 to $200 solo, $150 to $300 boutique, and $300 to $600 for Big Four firms. Enterprise transformation runs $75,000 to $200,000+ in year one, which is a different buyer entirely. These ranges synthesize the public 2026 cost guides. The one rate card you can read in full before contacting anyone is the c0nsl pricing section, covered near the end of this page.
The 2026 market, as one table
Every cost guide online presents this as paragraphs of caveats. It is cleaner as a grid. The ranges below are what the public 2026 guides converge on once you strip out the consulting language. Read the row that matches the work you actually need, and ignore the rows above it that exist mostly to make the middle look reasonable.
| Engagement | Typical 2026 range | What it buys | Notes |
|---|---|---|---|
| Paid consult or readiness audit | $75 to $300 | A scoped conversation, a written shortlist of automations, tool and model picks. Many firms run a free discovery call instead and recover the time in the quote. | Solo operators and small practices. Larger firms rarely charge for a first call because the call is a sales asset. |
| Hourly | $80 to $200 solo, $150 to $300 boutique, $300 to $600 Big Four | Open-ended work billed by time. Good for genuinely unknown scope, bad once the scope is known. | Common default. Risky for the buyer because the meter never stops and there is no fixed deliverable. |
| Single workflow build | $500 to $5,000 fixed | One automation shipped, tested, and documented: a support macro, a webhook pipeline, a report that replaces a recurring spreadsheet job. | Solo engineers and boutiques working fixed scope. The clearest unit to compare across vendors. |
| Custom multi-step system | $2,000 to $25,000 fixed | A real system: agents, a data pipeline, an escalation path, an audit log, and usually a few months of support folded in. | Solo to mid-size. This is where most small-business AI money actually goes. |
| Full project | $10,000+ fixed, quoted after paid discovery | Multi-system architecture, several integrations, team training, and a defined maintenance window. | Boutiques and up. Reputable shops will not quote this cold. |
| Monthly retainer | $1,000 to $8,000/mo for SMBs, $10,000 to $25,000+/mo for fractional leadership | Ongoing ownership: upgrades, model swaps, security reviews, the occasional middle-of-the-night fix. | Offered after a shipped build, or as a fractional executive arrangement. |
| Enterprise transformation | $75,000 to $200,000+ in year one | Organization-wide rollout, data infrastructure, custom training, four to twelve months of execution. | Big Four and large boutiques. This is not a small-business number, and it should not anchor yours. |
One thing the table makes obvious: the spread inside a single row is often wider than the gap between rows. A custom system quoted at $2,000 and one quoted at $25,000 are nominally the same line item. The next two sections are about why that spread is allowed to stay that wide.
Why the number is never on the page
Search this topic and you get the same shape every time: a tidy explanation of pricing models, a range so wide it commits to nothing, and a button that says book a free discovery call. The range is wide on purpose, and the call is not a formality. The call is the pricing instrument.
When a price is set after a conversation rather than before it, the conversation gets to do work. It can read your revenue, how urgent the problem feels, whether you have a deadline, and whether you are holding other quotes. A figure assembled with all of that in hand lands close to the most you would have paid without walking away. That is ordinary price discrimination. It is legal, it is everywhere in professional services, and it is not a scandal. But it is the reason a straight answer to a pricing question is structurally unavailable from the people who most want your business.
It also explains the second pattern you will have noticed if you went looking: a large share of the loudest content about automation pricing is not aimed at buyers at all. It teaches other people how to start an automation agency and how to price their own work. A buyer reading that material is reading a manual written for the person on the other side of the table. None of it is wrong, exactly. It is just not for you.
A consultant who posts a fixed rate card gives the discovery-call lever up deliberately. They cannot price-discriminate, they cannot anchor you in the room, and they invite you to compare them against the next consultant before you have spoken to either. That is a real cost to the seller. Most decide it is not worth paying. A few decide the trust it buys is worth more than the margin it gives up.
Hidden-rate quote vs published rate card
The two approaches are not just different prices. They are different buying experiences, and they put the work of comparison on different people. Here is the same engagement run through both.
| Feature | Hidden-rate quote | Published rate card |
|---|---|---|
| First contact | A free discovery call, booked before any number exists | A rate card already on the page; the call is for scoping the work, not for discovering the price |
| How the price gets set | The call reads your budget, your urgency, and your alternatives, then a number is built to fit | The same posted tiers apply to everyone, visible before you ever reach out |
| What the first call optimizes for | Anchoring you to the highest figure you will tolerate without walking | Naming three concrete automations and an hours-saved estimate you can act on |
| Comparing two consultants | Impossible until both have separately quoted you, days apart | Open two tier pages in two browser tabs and read them side by side in a minute |
| Scope creep mid-build | Re-priced verbally, often after the extra work has already started | Re-quoted in writing before any additional work begins |
| Who you are actually buying from | Often an agency brand, a reseller, or a course operator with a delivery team attached | A named senior engineer who takes the call and ships the code |
Neither column is fraud. The left column is simply the default in professional services, and the right column is a deliberate, slightly costly choice. Knowing which one you are dealing with on the first email tells you most of what you need to know about the next thirty days.
Four things that actually move your quote
If you want to sanity-check a number before you ever get on a call, estimate it yourself from these four drivers. They are listed roughly in order of how hard they push the figure. A consultant who cannot map a quote back to these four is quoting a feeling.
1. Scope shape
The single largest multiplier. One workflow, a multi-step system, and an organization-wide rollout are three different orders of magnitude, not three points on a line. Most small-business work is a single workflow or a small system. If a quote sounds like a transformation, the scope has drifted, and the price drifted with it.
2. Integration count
Every external system the automation has to touch adds authentication, error handling, retries, and testing. A pipeline that reads one inbox is cheap. The same pipeline wired to a CRM, a payment processor, and a calendar is a different job.
3. Data sensitivity
Clinic, legal, or finance data that genuinely cannot leave the building means local or private inference rather than a hosted API call. That is more engineering, and it is a legitimate reason for a higher number. It is also a reason to ask exactly where your data goes.
4. Who owns it after delivery
A build handed over with no retainer is cheaper today and tends to rot as models, APIs, and prices shift underneath it. A build with ongoing ownership costs more but stays alive. Decide which you are buying before you compare two quotes, because a build-only price and a build-plus-retainer price are not the same product.
What a published rate card looks like
Here is the part no competing pricing guide can copy, because it is not a market estimate. It is a real, posted card. The five tiers below are hard-coded in the c0nsl homepage source, in the PricingSection block of src/app/page.tsx, and they render live at c0nsl.com under section §07. You can read every number before you send a single email.
TIER-01
Small Integration
One workflow, shipped and documented. A webhook piece, a support macro, a report that replaces a recurring spreadsheet job.
$500 to $2K
fixed scope
TIER-02
Custom System
A real system, not a prompt. Local or hosted inference, a data pipeline, an escalation path, an auditable log.
$2K to $10K+
fixed scope
TIER-03
Full Project
Multi-system build for operators with a defined outcome and the budget to match. Never quoted cold.
$10K+
fixed scope, quoted after paid discovery
TIER-04
Retainer
Ongoing ownership of the AI surface you already shipped. Upgrades, model swaps, security reviews. Only opened after a shipped project.
$1K to $5K / mo
recurring, three month minimum
The anchor fact for this whole page is that those five lines exist in public, attached to a named senior engineer with fifteen years of shipped cross-platform work, priced deliberately below the $30,000 to $50,000 strategy-consultant route. Everything else in this guide is market math you can re-derive from any cost guide. This card is the one thing you cannot get from the pages that route you to a discovery call. No hourly billing, no retainer-first, no course upsell, and any scope change is re-quoted in writing before extra work starts.
Before any money moves, get these in writing
Whether you go with a posted rate card or a quote from a discovery call, the protection is the same: a written scope. Run this list against any proposal before you wire a deposit. A consultant worth hiring will not flinch at a single line of it.
Quote review checklist
- The deliverable is named as a specific system or workflow, not as 'an AI solution' or 'AI transformation'
- There is a fixed price or a fixed not-to-exceed ceiling, not an open hourly meter
- Scope changes are handled by a written re-quote before any extra work begins
- Model and infrastructure costs (Anthropic or OpenAI API spend) are a separate line from the consulting fee
- The escalation path is defined: which slice of the workflow stays with a human and never touches the AI
- Ownership of the code and the accounts after delivery is stated explicitly
- A named person will take the call and ship the work, not an account manager fronting a delivery team
Get a fixed-scope number, not a range
Bring the workflow eating the most hours. The consult comes back with three named automations, an hours-saved estimate, and a fixed-scope quote, or an honest note if AI is not the right tool.
More on pricing and scoping
Keep reading
Practice
Scaling solo AI consulting: the four ceilings nobody names
Why a published rate card and a solo practice fit together, and where scaling actually helps.
Read
Build vs buy
SMB AI agent pricing for Shopify: rent, build, or break even
Per-ticket math on the support-agent products versus a custom build, with real breakeven volumes.
Read
Scoping
AI automation for SMBs: the 20% to never automate first
The scoping framework behind every quote: which work is safe to automate and which has to stay human.
Read
Frequently asked questions
What does AI automation consulting actually cost in 2026?
For small-business scale work, a paid consult or readiness audit runs roughly $75 to $300, a single-workflow build runs $500 to $5,000 on a fixed price, a custom multi-step system runs $2,000 to $25,000 fixed, and an ongoing retainer runs $1,000 to $8,000 a month. Hourly engagements sit at $80 to $200 for solo operators, $150 to $300 for boutiques, and $300 to $600 for Big Four firms. Enterprise transformation programs run $75,000 to $200,000 or more in year one, but that is a different buyer and should not anchor a small-business budget. These figures synthesize the public 2026 cost guides; the one place you can see an exact, fixed, posted rate card is the c0nsl pricing section at c0nsl.com.
Why won't AI automation consultants tell me their price without a call?
Because the call is itself a pricing instrument. When the number is set after a conversation rather than before, the conversation can read your budget, your urgency, and whether you have other quotes, and the figure can be built to fit what you will tolerate. That is ordinary price discrimination, and it is legal and common. It is not malicious, but it does mean the headline ranges you find online stay deliberately wide so that no single number can be held against the seller. A consultant who posts a fixed rate card gives that lever up on purpose. Most do not.
Is a paid $75 consult worth it when other firms offer a free discovery call?
A free discovery call is a sales meeting; its job is to move you toward a proposal. A paid consult changes the incentive: the engineer is being paid to give you usable output in the room, which on a c0nsl consult means three named automations ranked by hours saved per week, tool and model recommendations, and an honest note if AI is the wrong tool. The $75 is refunded if that output does not materialize. The free call costs you nothing in dollars and an hour in time; the paid call costs $75 and is structured to send you away with something you can act on even if you never hire anyone.
What is a fair price for automating one workflow at a small business?
A single, well-defined workflow (a support-ticket triage macro, an inbound form router, a scheduled report that replaces a spreadsheet) is a fixed-scope job and should be quoted as one. Fair fixed pricing for that runs roughly $500 to $5,000 depending on how many systems it touches and how sensitive the data is. If a consultant will only quote it hourly with no ceiling, treat that as a signal that the scope has not actually been pinned down yet. A workflow you can describe in two sentences can be priced as a fixed number.
Hourly, fixed-scope, or retainer: which model should a small business pick?
Pick fixed-scope for anything you can describe concretely, because it moves the risk of an estimate being wrong onto the consultant, where it belongs. Use hourly only for genuinely open-ended discovery where neither side can yet name the deliverable, and even then ask for a not-to-exceed ceiling. Take a retainer only after a real system has shipped, because a retainer on day one is paying for maintenance of something that does not exist yet. The c0nsl card follows exactly that order: a flat consult, then fixed-scope build tiers, then a retainer that only opens after a shipped project.
How do I tell a real automation quote from an inflated one?
Three checks. First, the deliverable is named as a specific system or workflow, not as 'an AI solution' or 'AI transformation.' Second, model and infrastructure costs (the Anthropic or OpenAI API spend) are listed as a separate line from the consulting fee, because bundling them hides both. Third, the quote says what happens when scope changes, and the honest answer is a written re-quote before any extra work starts. A quote that fails all three is not necessarily dishonest, but it is unfinished, and you are within your rights to send it back for a real one.
What pushes a quote higher: what are the cost drivers?
Four things, roughly in order of impact. Scope shape is the biggest: one workflow, a multi-step system, and an organization-wide rollout are three different orders of magnitude. Integration count is next, because every external system the automation touches adds authentication, error handling, and testing. Data sensitivity is third: clinic, legal, or finance data that cannot leave the building means local or private inference, which is more engineering than a hosted API call. Ongoing ownership is fourth: a build with no retainer is cheaper upfront but tends to rot as models and APIs change underneath it.
Comments (••)
Leave a comment to see what others are saying.Public and anonymous. No signup.