Splunk and AI consulting
Get more out of the Splunk you already pay for
Most organisations I meet have the data in Splunk and a licence that grows every year. What they are short of is the time and the architecture to turn it into faster investigations, clearer root causes and work that runs without someone having to start it. That is the work I take on.
I am a Splunk Enterprise Certified Architect with thirty years in network and performance engineering, and I built Adjutant AI, a governed AI platform that runs inside Splunk. What I learned building it goes into your environment, whether or not you ever use my software.
Architect work, not day-to-day administration. A small number of engagements at a time.

When organisations call me in
The questions differ from one organisation to the next. The situations behind them repeat. If one of these sounds familiar, we should talk.
The licence grows, the value does not
Ingest and cost go up every year, but the same handful of dashboards get used. Nobody has had the time to ask which data earns its place and which use cases are missing.
The same investigation, twenty times a day
Analysts repeat the same pivots for the same type of notable or ticket. It is skilled work done by hand, and it is the first thing that should run on its own.
The AI pilot stalls at the security review
Someone built a promising AI demo on Splunk data. Then came the questions: where does the data go, whose permissions does it use, who audits it. The project stopped there.
Red, and nobody knows why
Services turn red, incidents come back, the platform itself slows down. Hours go into finding out whether the problem is the system, the feed or Splunk.
Typical engagements
Each one starts from a result you can check, not from a number of days. These are the shapes the work usually takes. The scope is always set against your situation.
Splunk value and AI readiness scan
Where your Splunk licence pays off, where it does not, and the three use cases worth doing first. The usual starting point.
- Which data is ingested, which is used, and what it costs
- Whether the data is reliable enough to automate on
- Three use cases, each with effort, owner and expected return
- A short report and a session with your team and management
Investigation automation
One type of security or operations investigation taken from manual work to a triggered, repeatable run, with time and quality measured before and after.
- One investigation type: an ES notable, a phishing report, a suspicious login or a recurring ITSI episode
- The steps your analysts take today, and the data and tools behind them
- A governed playbook that starts on its own and leaves evidence an analyst can check
- Time to verdict and quality, measured before and after
Root cause analysis
An evidence-backed answer to a problem that keeps coming back, in your services or in the Splunk platform itself, plus the check that catches it next time.
- Recurring incidents and services that turn red without a clear cause
- The platform itself: skipped searches, indexing delay, search head load
- Broken feeds hiding behind alerts: changed formats, timestamps, parsing
- A written root cause, the fix, and monitoring that keeps it fixed
Connecting tools and AI to Splunk
Splunk working with the systems around it, and your corporate AI chat answering Splunk questions through defined scenarios instead of open access.
- ServiceNow and CMDB, Cribl Stream, Edge Processor, TrackMe and your own APIs
- MCP design: which scenarios to expose, under which roles, with which audit trail
- Model and hosting choices, including fully offline models for sensitive data
- Integrations with an owner, replay and drift checks, not scripts nobody maintains
Not on this list? These four cover most requests. An ITSI service design, a data onboarding strategy or an architecture review before a bigger change fit the same way of working. Describe the problem and I will tell you whether I am the right person for it.
How I work
Small enough to start next month, concrete enough to measure.
Start with one problem
We pick one problem with a visible cost in time, risk or licence. Not a programme and not a roadmap deck.
Agree how we measure it
Before anything is built, we write down what better looks like: time to verdict, hours saved, incidents avoided, data actually used.
Build it in your environment
The work runs in your Splunk, under your roles and on your data, so what we measure is real.
Hand it over
Your team gets the documentation and the understanding to run and change it. Keeping it working should not depend on me.
Independent advice, including about my own software
I build Adjutant AI and I will tell you when it fits. I will also tell you when it does not. Splunk’s own AI Assistant, the Splunk MCP Server, SOAR or a search you already have can be the better answer, and then that is what I recommend.
- Use what you have first. Splunk and the tools you already pay for come before anything new.
- No product as a condition. Consulting is paid as consulting. Buying Adjutant AI is never part of the deal.
- Alongside your partner. If you work with a Splunk partner, I work with them, not around them. They know your estate; I add a specific depth.
- Honest about what is possible. If something cannot be done yet, or is not worth doing, you hear that from me before you pay for it.
Why me
Plenty of people can advise on Splunk and AI. Fewer have built a governed AI platform inside Splunk and run it in regulated environments.
Built, not only advised
Adjutant AI connects Splunk to ServiceNow, Cribl, TrackMe and MCP, runs investigations on triggers and computes ITSI root cause in code. The patterns I recommend are ones I have built and run myself.
Splunk’s highest certifications
Splunk Enterprise Certified Architect and Core Certified Consultant. I learn a technology down to its concepts before I advise on it.
Regulated environments
Years as an ITSI specialist in banking, insurance and automotive, and co-presenter with Rabobank at Splunk .conf2017.
Thirty years of asking why
Network engineering from 1994, performance engineering at a major Dutch bank from 1997. Root cause analysis is where my career started.
The bottleneck is rarely Splunk itself. It is the distance between what people need to know and what they can get out of it.
Is this a fit?
A good fit
- You run Splunk Enterprise or Splunk Cloud and want more from it
- There is a concrete problem with a visible cost in time, risk or licence
- Someone on your side owns the result and can make decisions
- A fixed-scope project or a longer architect role, as long as the goal is clear
Not a fit
- Day-to-day administration, upgrades or working a ticket queue
- Filling a seat without a goal we can measure
- An AI demo meant to impress rather than to run in production
- Work where the answer has already been decided
For daily administration I am glad to point you to a good Splunk partner.
One person, on purpose. I take on a small number of engagements at a time, next to building Adjutant AI. The person you talk to is the person who does the work, from the first conversation to the handover.
Bring me one problem
Describe a situation where Splunk should be helping more than it does. I will tell you honestly whether I can help, what I would look at first and roughly what it would take. The first conversation is free.
