Managed AI service, consultant, or software: which fits your business?
The main difference is who sets up the automation and keeps it working. For document intake, your team might configure and run software, hire a consultant for advice or a build, or use a managed service for ongoing operation. Providers can combine these roles. Compare each proposal's responsibilities for setup, maintenance and human review, rather than relying on the label alone.
Who does what in each model
| Question | Software you buy | A consultant | A managed service |
|---|---|---|---|
| Who sets it up | Your team, with the vendor's tools and documentation | The consultant, as far as the engagement covers | The provider, with your team explaining the process |
| Who keeps it running | Your team | Your team, unless you agree ongoing support | The provider, as part of the monthly service |
| How you pay | A subscription or license | Set by the engagement: ask whether it is a project, hourly or retainer fee | A recurring fee for the agreed service |
These are examples; providers may combine them. Your contract sets the responsibilities.
When software alone is enough
If the work fits a tool's standard features, and someone on your team has time to configure it and look after it, software may be all you need. Check what your current software can already do before you buy anything new. The useful starting point may be a feature you already have.
When a consultant fits
A consultant fits when you want a plan, an independent assessment or a one-time build, and you have people who will run the result afterwards. If what you want is a list of tools to buy, a consultant suits you better than SubAI: tool recommendations alone are outside our service.
When a managed service fits
A managed service can fit recurring work that affects customer service, staff capacity or revenue. Look for a provider that builds and maintains the agreed automation, reports how it performs and takes responsibility for fixing problems within the agreed scope.
What stays with you, whichever you choose
Your team still explains the process and makes the business decisions. Agree who provides access, who reviews the work and who approves messages or other actions. Confirm in the contract who owns the accounts, data and documentation, and how you can retrieve them if the engagement ends.
Questions to ask any provider
- Who maintains the automation when a connected system or form changes?
- What result will you measure, how will you test it, and when does billing begin?
- Which steps run automatically, and which go to a person for review?
- Who owns the accounts, licenses and documentation if you leave?
- What happens if the work misses the agreed test?
- How will you see what ran, how it performed and what needs attention?
- What does it cost after launch, and what would change the price?
How SubAI answers them
We start with a discovery call. If the opportunity looks worthwhile, a $2,000 diagnostic maps the work, estimates current costs using your figures and recommends where to begin.
For a first automation build, we agree on the success test and monthly fee in writing. No implementation fee. The build and 30-day proving period are unbilled; monthly billing starts when the agreed test is met.
You own your accounts and licenses. Standard engagements include weekly reports, or monthly if you prefer. Separately scoped data connections and reporting have their own operating commitments and billing terms. See pricing for options and how we assess fit.