Questions we get, including the awkward ones.

What does After AI Consulting actually do?

Enterprise AI consulting across nine business functions and nine industries, from strategy through build and into managed operations. The service areas are AI strategy and governance, engineering enablement, legacy code modernization, AI automation and integration, IT and MSP operations, and cloud, data and platform. The work is operational rather than advisory: we write the policies, build the systems, and do the migrations alongside your teams.

How is this different from a large consultancy?

Scale and who touches the keyboard. A large firm brings a partner to the pitch and a team of analysts to the delivery. Here the people who run the assessment are the people who do the implementation, and the deliverable is a working thing rather than a recommendation. The trade is real: we cannot staff a twelve-country rollout, and we will say so rather than take the engagement.

How much does it cost?

We do not publish prices, because every engagement is scoped to the systems and teams involved. For calibration against the market: published 2026 pricing guides put a mid-market AI readiness assessment at roughly $10,000 to $25,000 over three to six weeks, implementation work starting around $40,000, and fractional AI leadership at $5,000 to $10,000 a month. Check any quote, including ours, against those ranges.

Is any of the work done outside the United States?

No. Every engagement is delivered inside the United States. The people on your engagement are our own employees, based in the US, not subcontractors and not an offshore delivery centre, and the work is performed here rather than handed to another timezone overnight. Our own suppliers are US companies, so hiring us does not add a foreign entity to your subprocessor list. We do this because most of our clients are vendors to somebody larger, and their customers’ vendor-risk questionnaires ask where work is performed and who can reach the data.

Who is on the team?

A team rather than one person, though names are not on the site yet because several of us are still wrapping up prior commitments. Combined, the team has 30 years of systems and software architecture, 30 years of database administration, 20 years of QA and application testing, and 20 years working alongside managed service providers, and has used frontier language models in client work since ChatGPT was released in late 2022.

Do you resell any AI tools?

We hold partner relationships with some vendors and can resell where that is the cheaper route for you. Where a recommendation carries a commercial interest for us, we say so inside the recommendation rather than in a footnote. If you would rather we did not resell on your engagement, say so at scoping and we will quote it that way.

Will AI let us reduce headcount?

The evidence does not support planning on it, and we will not sell it that way. PwC’s 2026 analysis found the most AI-exposed companies grew headcount 52% between 2018 and 2025 against 36% for the least exposed. The one large field study of an AI assistant found attrition 8.6% lower among workers who had it. What changes is what people do, and the organizations that keep their people are the ones that decide what the new work is before the tools arrive.

We already started and it is going badly. Can you help?

Usually, and this is a common starting point. The failures we see repeat: tools deployed with no policy, no measurement baseline so nobody can tell whether it worked, agents given more authority than anyone intended, or a project stalled on data that was never assessed. The assessment identifies which of those it is, and it is cheaper than continuing.

Do we need an assessment before you will do implementation work?

Yes, unless you have done something equivalent in the last six months. Every implementation we have seen go wrong started without one. It is also how we decide whether to take the engagement, which cuts both ways.

What if the assessment finds nothing worth doing?

Then the report says so, with reasons, and it is the same fee. That happens, and it is useful to have in writing when the board asks again next quarter. Sometimes the honest recommendation is to fix your data with your own team first and revisit AI in two quarters.

Do you work with managed service providers?

Yes, and it is the sharpest specialism here. Two things: getting AI into the MSP’s own service delivery (triage, documentation, runbooks, monitoring), and building an AI consulting and governance practice the MSP can sell to its own clients under its own name. Kaseya’s 2026 survey found 48% of MSPs rank AI as the top client need while only 13% earn meaningful revenue from it; that gap is the work.

Can you run AI models on our own hardware?

Yes. Open-weight models on your servers or private cloud, with the same tooling built against them, for cases where data cannot leave the building. They are less capable than frontier models and we will tell you precisely where that shows up rather than letting you discover it in production.

Do you take equity instead of fees?

No. Fees only.

What do you need from us during an engagement?

Access to the people who do the work rather than only the people who manage it, read access to the systems in scope, and one person on your side who can get answers. About four hours of interviews across departments for an assessment. No slide decks.

What do you refuse to work on?

Three things. EU AI Act conformity work for high-risk systems, which belongs with specialist regulatory counsel. Anything that makes an employment, credit, or clinical decision without a documented human decision behind it, because that is a regulatory exposure rather than an efficiency. And individual developer productivity scoring from AI telemetry, which measures activity rather than outcome and costs you the team’s trust.

Where are you located?

Remote, working across the United States.

Several of these have their own page with the evidence behind them: what AI consulting costs, whether AI reduces headcount, and the rest of the question pages.

Ask us. A person reads every message and replies.

Talk to us