Escalations from tier one to tier two
The main sign that techs are solving more on their own.
IT managed service providers · Internal knowledge assistant
Help newer techs solve tickets without interrupting your senior engineers
Every MSP has two or three engineers who know why a client's VPN is configured the way it is, which line-of-business app breaks after updates, and who to call at the ISP. Newer techs interrupt them all day. Your documentation platform holds the facts, but finding the right article across hundreds of clients takes longer than asking in Teams.
How it often works today
- Tier-one techs escalate tickets they could solve because they cannot find the client-specific procedure in IT Glue or Hudu fast enough.
- Senior engineers lose focus to a steady stream of Teams messages asking about client setups they configured years ago.
- Resolved ticket notes contain valuable fixes, but nobody searches old tickets, so the same problem gets diagnosed from scratch each time.
- New hires take months to become productive because tribal knowledge is passed on through shadowing rather than written down.
We would set up an assistant that answers tech questions from your own runbooks, client documentation and resolved tickets, and cites where each answer came from. It respects client permissions, never surfaces stored credentials, and shows when the underlying article is stale so someone can fix it.
The improved workflow
Audit what is documented
Review the documentation platform for coverage, ownership and staleness, and identify the twenty questions techs ask most often in Teams and in escalated tickets.
Connect approved sources
Index runbooks, client procedures and resolved ticket summaries, excluding password fields, secure notes, credential attachments and anything a client has flagged confidential.
Answer with citations
Techs ask in plain language within Teams or the PSA. Answers link to the source article or ticket so the tech can verify before acting on a client system.
Feed gaps back into documentation
Questions the assistant cannot answer, or answers marked wrong, create a documentation task for the right engineer, so coverage improves every week.
How we deliver this: AI automation for small business or fractional CTO leadership, scoped in writing and run over 90 days. See pricing or the sample scope.
What we would measure
Interruptions to senior engineers
Measured by sampling Teams questions before and after the pilot.
Answers marked wrong or unhelpful
Keeps quality visible and prioritizes documentation fixes.
Time to productivity for new hires
A longer-term measure of whether knowledge is really shared.
Where people stay in control
- Stored credentials, MFA seeds and secure notes are excluded from indexing entirely; the assistant points to IT Glue or Hudu rather than reproducing secrets.
- Answers about changing a client's production system are treated as guidance; techs follow the change process and get approval for risky work.
- Access follows existing roles, so a tech only sees documentation for clients and areas they are permitted to work on.
Example
A Friday afternoon line-of-business app issue
For example, in an illustrative case a tier-one tech gets a ticket that a dental client's imaging software cannot reach its database after Windows updates. Instead of messaging the senior engineer, the tech asks the assistant. It returns the client's runbook step for restarting a specific service in a set order, plus a resolved ticket from eight months ago with the same symptom. The tech follows the documented steps, resolves it, and marks the answer helpful.
Tools typically involved
What the first 30 days look like
In the first 30 days we would collect a sample of escalations and Teams questions, review documentation coverage for your top clients, decide which sources are safe to index, and agree how answers will be checked. A small group of techs would test it first.
Common questions
Is it safe to point AI at our documentation platform?
Only with careful scoping. We would never index password fields or secure notes, and we would prefer tools whose data stays within your Microsoft 365 tenant or under clear processing terms. A typical first step is indexing runbooks and procedures only, then adding ticket history once access controls are proven.
Our documentation is out of date. Is this still worth it?
That is common, and the assistant helps expose it. Wrong or missing answers turn into specific documentation tasks, which is more useful than a general instruction to update the docs. We might spend part of the first month fixing the most-used procedures before the assistant goes live with the team.
Would Microsoft Copilot do this already?
Copilot can search SharePoint and Teams content you already have permission to see, which may cover part of the need. It does not automatically know your PSA ticket history or documentation platform. We would test what Copilot handles well before adding anything, since using a license you already pay for is often the cheaper answer.
The business case
Does it pay for itself?
Illustrative defaults for a managed service provider. Replace them with your numbers.
What this workflow costs today
a month in staff time, before any lost sales
Monthly fee + assumed tools +
Break-even: new clients a month (at first-year value)
Or: staff hours recovered and redeployed
Break-even additional revenue
Value of +1 point of win rate
First 90 days, service fee
Worth a closer lookTest the assumptions
Illustration only, not a forecast or guarantee. Recovered staff time becomes a cash saving only if spending actually declines or the time goes into paid work. Tool costs are an assumption; we replace every number here with yours before recommending a scope.
Internal knowledge assistant in other industries
Accounting and CPA firmsArchitecture firmsConsulting and professional services firmsEngineering firmsLaw firmsSmall manufacturersNonprofit organizations
Internal knowledge assistant: full guide
Related guides
Getting staff to adopt new toolsMicrosoft Copilot for small businessCybersecurity basics to fix before adopting AIWriting standard operating procedures with AI, and keeping them true
Terms used here
Next step
Bring one expensive workflow. Leave with a clear next step.
In 30 minutes we look at one recurring workflow, what it costs you today and whether there is a credible case to improve it. If there is not, we will say so.
Book a 30-minute workflow review
Not ready to talk? Take the 3-minute check
30 minutes with Phil or Martin. No technical brief, no pitch deck. Phil or Martin replies personally, usually within one business day.