Process and AI consultancy
Lean process improvement and AI automation, built to be relied on.
We work with organisations that run on high-volume manual work. We simplify a process before we automate it. Your teams spend fewer hours on manual work and make fewer errors, and your customers get faster, more consistent service.
- 9years in process improvement and automation
- 4regulated sectors
- 30,000+hours of manual work saved to date
- €1.1m+in savings delivered
Figures from our founder's nine years of delivery.
What we do
-
Process improvement
We map how the work runs, measure it, remove non-value-add steps and standardise what remains, based on Lean Six Sigma.
-
Workflow automation
We build automations that run tasks and whole workflows across the systems you already have, from approvals and handoffs to reports, reconciliations, alerts and follow-ups.
-
AI automation and agents
We design and build AI agents that read emails, documents and records, decide what should happen next, and complete multi-step tasks in your systems.
-
AI reliability
AI will sometimes be wrong, so it is not right for every task. We work out with you what rules, AI and people should each do, and send anything uncertain to a person.
We work across customer service, finance, operations, HR and back office. Our deepest experience is in contact centres.
A sample of past projects.
These four are replayed on sample data, with the clients anonymised. Each started with a problem a client brought to us.
Past project, sample data
Also delivered
- Conversational AI flows
- Contact centre call tagging
- Accounts payable automation
- Compliance reporting
- KPI dashboards
- Data migration
Bring us a process from any function. The method is the same.
AI reliability
We plan for the AI being wrong.
AI is not the right tool for every task. For each one we ask what a mistake would cost, then recommend who should do it. You make the final decision.
| If a mistake is | The work goes to |
|---|---|
| Impossible, because the answer is certain | Fixed rules |
| Cheap, visible and easy to undo | AI alone |
| More costly | AI prepares it, a person approves |
| Costly, irreversible or regulated | A person |
If one mistake in a hundred would be costly, the task stays with a person, however accurate the AI is.
Results
Numbers from client projects.
Most of this work is under NDA, so clients stay anonymous. Figures marked est. are estimates.
-
AI email triage · Health insurance
- 53% to 96% accuracy
- 6,000+ hrs a year saved
- €175k a year net benefit
One stalled AI project, taken live.
-
Revenue leakage · Telecoms
€200k+/yr
Annual recurring revenue recovered
New customers stopped at the document upload and never came back. Automated reminders now get them to finish.
-
Time savings · Telecoms
1,600 hrs/yr
Agent time saved
Agents no longer warn customers about usage spikes by hand. Alerts go out automatically, on each customer's own settings.
-
Revenue leakage · Restoration
$30k/yr
Additional revenue
Drying equipment that sat idle is now tracked, matched to jobs and billed.
-
Time savings · Restoration
3 hrs est. to 20 min est.
Daily job review
A project manager gets one ranked list of the day's actions instead of opening every job file.
-
All case studies
Eight case studies
Each one covers the problem, what we built, the steps and a note on the figures.
How we work
Measure first, then automate.
Every project runs the same seven steps. You get a baseline before anything changes and the same numbers again after go-live, so you judge the result on your own figures.
-
01Your current state
Understand
How the work is done today, who does it and what it costs.
You get:Process map of how work runs today
-
02Problems and opportunities
Find the gaps
Where the process breaks, and which fixes would pay back most.
You get:Ranked list of opportunities
-
03Before
Measure the baseline
The numbers you'll judge us on, taken before anything changes.
You get:Baseline scorecard of hours, errors and revenue at risk
-
04The process
Simplify
Cut the steps that shouldn't exist, so we don't automate waste.
You get:Redesigned process
-
05The solution
Design and deliver
Built into the tools your team already uses, with a person checking anything that needs judgement.
You get:Working automation and handover guide
-
06After
Measure the outcome
The same numbers, measured again after go-live.
You get:Before and after report
-
07
Keep improving
Your staff flag anything the automation gets wrong. We find the cause, fix it and re-test before release.
You get:An accuracy report and a log of changes
It's Lean Six Sigma's DMAIC method, adapted for work that now includes AI.
In practice
One tax report, from weekly to daily.
A telecoms operator's purchase VAT report came out wrong every time it was downloaded. Two analysts fixed it by hand, which is why it only ran weekly. Every missed run was another chance for a tax error to get through.
Read the full case- 01 Extract Pulled from the finance system daily.
- 02 Clean Fixes rows, removes duplicates.
- 03 Decide Tax rules label each line include, exclude or human review.
- 04 Reconcile Matched to the general ledger before anything is released.
- 3 hrs to 0 hrsmanual work per run
- Weekly to Dailytax data checked five times as often
- 100%transactions traced to the rule that decided them
Experience
Five sectors, four of them regulated.
- Health insurance
- Telecoms
- Financial services
- Aviation leasing
- Restoration services
About
Krishna Saish
Krishna founded Focal IT Consulting and works from Limerick, Ireland. He holds a first-class Master's from University College Cork, a Lean Six Sigma Green Belt and the Agile Certified Scrum Product Owner credential.
More about KrishnaQuestions
Before you book
What kind of results can I expect?
It depends on the process, which is why we measure it first. We count hours, errors and revenue at risk, then count them again after go-live. You judge the result on your own numbers.
Do I need to buy new software?
Usually not. We build into the systems your team already uses, and what we build is yours.
Is it safe to use AI with my customer data?
We've delivered automation in health insurance, financial services and telecoms, where data protection sign-off is part of the job. Every design sends uncertain cases to a person, and sensitive data stays out of anything that doesn't need it.
How do you charge?
Fixed price per scoped process, agreed before work starts.
How do we get started?
Book a call and bring one process. If it's worth automating, we'll send a scoped proposal with a fixed price. If it isn't, we'll tell you.
Bring us one process.
Pick the process your team complains about most. Start with a 30-minute call.
No obligation.