What is the real difference between the two?
An in-house team is a capability. You are buying people who will be around next year, who learn your firm's quirks, and who can pick up the next job and the one after that. A consultant is a delivery. You are buying a defined result, within a defined period, from people who already know how to produce it and who then leave.
Neither is better in the abstract. The right answer turns on how much AI work you actually have, how soon you need something to change and whether you can find and keep the right person.
How do they compare side by side?
| AI consultant | In-house team | |
|---|---|---|
| Cost structure | Paid per engagement, either a fixed fee or time and materials. Stops when the work stops. | Salary, employer costs, recruitment, tools and management time. Ongoing whether or not there is a project that month. |
| Time to value | Weeks, if the scope is one clear process. | Months. Recruitment, onboarding and learning the firm come before any build. |
| Risk | Sits with the provider if the fee is fixed and tied to delivery. Sits with you on day rates. | Sits with you. A poor hire, or a good one who leaves, sets you back to the start. |
| Control | You set the goal; the provider decides how. Control over the result depends on the contract. | Full. Priorities can change week to week. |
| Knowledge | Arrives with experience of the method. Has to learn your firm quickly. | Knows your firm well over time. May have to learn the method from scratch. |
| What happens after | A handover. You keep what was built and the team is trained, if that was agreed. | The capability stays, as long as the people do. |
| Best for | One job, fixed quickly, with a clear finish line. | A long list of work and a plan to keep going for years. |
When is an in-house team the better choice?
Hire when you already know there is a long queue of work, not a single job. If the partners have agreed that AI will change how the firm delivers across several practice areas over the next few years, a permanent team makes sense, because the ongoing cost is spread across many projects and the knowledge stays in the building.
It also suits firms whose work is so specialised that an outsider would spend most of an engagement learning it. And it suits firms that already employ someone technical, such as a knowledge manager or a systems lead, who has appetite and time to take this on. In those cases you should not hire us. You should back your own person.
The honest caveat is the hiring itself. People who can both build working systems and understand how a fee earner spends a day are scarce. A firm of 50 people may struggle to offer the career path that keeps them.
When is a consultant the better choice?
Bring someone in when the problem is specific and urgent: a single process, such as proposal production or client reporting, is eating senior time now, and nobody inside has the hours to fix it. A consultant who has done this before can start mapping in the first week, rather than in the first quarter.
It also suits firms that are not yet sure how much AI work they have. One finished process tells you more about the size of the opportunity than any strategy paper, and it costs you an engagement rather than a permanent salary. If that first rebuild proves the point, you are in a far better position to write a job description.
How should you decide?
- Count the work. List the processes you would want changed in the next two years. One or two points to a consultant. A dozen, with more behind them, points to a hire.
- Check who you already have. An existing person with time and interest is worth more than either option. Give them the hours before you spend elsewhere.
- Be honest about urgency. If a process is costing a partner a day a week today, the months it takes to recruit are a real cost. Work it out with the unbillable hours calculator.
- Look at what is left behind. Whichever you choose, the result should run in your own tools and be understood by your own team. Ask any consultant exactly what you will own at the end.
Aldbry is the consultant option in its narrowest form: one process, rebuilt in 30 days, for a fixed fee, with the team trained and 30 days of support after go-live. If you have a long programme and the budget for a permanent team, that is probably the better investment. If you have one job that needs fixing first, the audit will show which one.