When an outside advisor adds value
An advisor is not a replacement for your team. They are most useful when you face a decision that is hard or expensive to reverse, and you lack the time, the specific experience or a perspective that is not tied to the project's history. When the question is simple and the team already knows the answer, outside advice usually just slows things down.
Pay particular attention to situations where the discussion has stalled. Two experienced people on the team hold opposing positions, a vendor proposes a solution that is hard to compare, or a proof of concept works in the demo and goes nowhere. In cases like these, an outside view can help break the question apart and bring it back to criteria you can actually test.
- A decision that is hard to reverse: choosing a platform, a data model or a key vendor.
- Professional disagreements the team cannot resolve on its own.
- A vendor quote or technical proposal that needs evaluating.
- A POC that works in the demo and is stuck on the way to production.
- An area that is new to the team, such as AI systems or cloud infrastructure, with no internal experience yet.
Frame a question that can be answered
A request like "tell us what to do with AI" is too broad. Frame the question around a decision: what needs to be decided, by when, and what happens if nothing is decided. For example: should we stay with the current vendor for the next two years or start a migration, and under what conditions.
Give the advisor the context they need: budget constraints, regulatory and privacy requirements, team structure and what has already been tried. Also state what is not up for discussion. Clear boundaries save hours of recommendations that could never be implemented in your environment.
What a good deliverable looks like
A consulting deliverable should hold up after the advisor has left. A good conversation is a start, but a recommendation that exists only verbally erodes quickly. Ask for a short written document that the team can read, challenge and come back to when the context changes.
A good recommendation also shows what was not chosen and why. It flags the assumptions it relies on, so you can tell when it no longer applies. If it contains no risks or trade-offs, something is probably missing.
- A written recommendation with the core rationale in one paragraph.
- Two or three alternatives that were considered, with the benefits and cost of each.
- Risks, assumptions and open points that need further checking.
- A roadmap with a small first step you can start on.
Avoid conflicts of interest
Ask the advisor directly whether they have a business relationship with vendors or tools they might recommend: referral fees, partnerships or a commitment to sell implementation hours. A relationship like that does not necessarily disqualify them, but you need to know about it up front so you can weigh the recommendation accordingly.
Where possible, separate the recommendation phase from the implementation phase. If the person recommending is also the one who will build, make sure the recommendation includes real alternatives and that you could hand it to another team or vendor without losing its substance.
How to tell whether it was worth it
Decide in advance what will count as a result: a decision that was made, a risk identified before the investment, or an option ruled out in time. Consulting is not measured in slides. It is measured by whether you made a better decision than you would have made without it, and that cannot always be fully proven.
Revisit the recommendation after a few months. Check which assumptions turned out to be right, what changed and what you would do differently. The answer will help you decide whether and when to bring in outside help again, and at what scope.