Consultancy, training, or in-house for AI search, which fits
The decision between hiring a consultant, training your team, or building in-house depends on three variables founders usually get wrong.
You have budget for AI search. You know it matters. The question is not whether to invest, it is how.
Do you hire a consultant, send your team to training, or build an in-house function?
The answer depends on three variables, and founders I work with in Jakarta, Singapore, and Kuala Lumpur usually get at least one of them wrong.
The three variables that decide the model
Speed. How fast do you need a working practice?
If you need a live roadmap, instrumented tracking, and an editorial pipeline running in 60 days, you do not have time to train a team from scratch. If you have six months and are hiring anyway, training might be cheaper and leave you with more durable capability.
Internal capacity. Do you have someone who can own this full-time?
Not "find two hours a week." Own it. If the answer is no, consultancy or in-house are your only real options. Training without an owner turns into Notion docs nobody reads.
What you actually need. A strategy, execution, or both?
In my consulting work I have seen founders pay for training when what they actually needed was someone to write the roadmap, set up schema, and instrument the tracking. I have also seen the opposite, a retainer when the founder had a sharp product manager ready to learn and run it themselves if someone would just show them the system once.
Most decision frameworks stop here. That is not enough. You also need to know what each model actually delivers, and what it assumes you already have.
What you get in each model
| Model | What you get | What you need to supply | Timescale | Handoff |
|---|---|---|---|---|
| **Consultancy** | Roadmap, audit, prioritized tasklist, monthly strategy calls | Someone internal to execute the roadmap (or you hire for it) | 2–4 weeks for audit + roadmap, then ongoing retainer if needed | You own it after the engagement ends |
| **Training** | Structured curriculum, live coaching, operating templates, Q&A access during cohort | A team member with time to learn and apply it, willingness to do the work | 4–12 weeks depending on format (workshop vs. cohort) | You run it solo from week one |
| **In-house / embedded** | Full execution (content, schema, tracking, reporting) OR a built-and-trained team you then own | Budget and a decision on whether you want execution or team-building mode | 90 days to first milestone, 6–12 months for team-building mode | Depends on mode. Execution is ongoing. Team-building mode ends when your hire is trained and running the practice. |
The table makes it look clean. The operational reality is messier.
Consultancy assumes you can execute a prioritized tasklist once someone hands it to you. If your writer is already booked at 110 percent, the roadmap will sit in a slide deck gathering dust.
Training assumes your team member has the appetite to learn a new discipline and apply it without a safety net. If that person is a founder juggling product, fundraising, and hiring, the course gets half-finished.
In-house embedded work assumes you trust the team enough to let them run the function or build your hire from scratch. If you want tight control over every heading and every schema edit, embedded execution will feel expensive. If you want to own the function long-term and are hiring for it anyway, team-building mode usually makes more sense than paying for ongoing execution.
The honest tradeoffs nobody puts in the marketing deck
Consultancy is faster but assumes you have execution capacity.
I can deliver a prioritized AI search roadmap in two weeks. I cannot make your product manager suddenly have 15 hours a week to write the content, edit the schema, and track the citations. If that capacity does not exist, the roadmap becomes expensive advice you never action.
When consultancy works well in the SME engagements I run, it is because the founder has already hired a marketer or a product person who is curious, capable, and has time. The consultant sets the strategy, the internal person runs it. Monthly calls keep it on track.
When it fails, it is because the founder thought strategy alone would move the number. It will not.
Training is cheaper and builds internal capability, but it is slower and requires discipline.
A four-week cohort costs a fraction of a six-month retainer. You leave with templates, a clear system, and enough knowledge to run AI search and SEO as one practice without ongoing help.
The tradeoff is that you have to do the work. In the cohorts I run, I have watched founders who finished every assignment build practices that are still running two years later. I have also seen founders drop out in week two because a product crisis ate their calendar.
If your team does not have the time or the discipline to finish a structured program, training will not fix that. It will just be another half-finished course.
In-house embedded execution is the most expensive and the fastest to a working practice.
You get a functioning search practice without hiring, without training, and without figuring out the instrumentation yourself. Someone senior shows up, audits your site, builds the schema, writes the content plan, sets up tracking, and reports on it monthly.
That model works when speed matters more than cost and you do not want to own the function internally yet. It stops working when the engagement drags on for 18 months because you never hired someone to take it over.
Team-building mode costs less over time if you are hiring anyway. We recruit the person, train them, run the practice together for six months, then hand it off. You end up with a trained hire and a working system. The tradeoff is that it takes longer and you need to be ready to recruit.
A decision tree you can actually use
Start here.
Do you have someone internal who can own AI search and SEO full-time, today?
- Yes, and they are capable but need a system and some coaching? → Training.
- Yes, but they are at capacity and you need strategy, not execution help? → Consultancy.
- No, and you are hiring for the role in the next three months? → In-house team-building mode.
- No, and you are not hiring? → Consultancy if you can execute the roadmap yourself, or in-house execution mode if you cannot.
How fast do you need a working practice?
- 60–90 days? → Consultancy or in-house execution.
- 4–6 months and you are okay learning as you go? → Training or in-house team-building.
What is the actual constraint?
- You do not know what to do? → Consultancy.
- You know what to do but nobody has time to do it? → In-house execution or hire for it.
- You want to own this internally long-term and have budget to build the function? → In-house team-building mode.
The wrong decision is not picking consultancy when you needed training. The wrong decision is picking any model without the internal capacity or discipline to make it work.
If you are still not sure which model fits, book a 30-minute call and we will figure it out. No pitch, just a clear recommendation based on where you are today and what you actually need to move the number.