Slow ERP and back-office reports
The month-end report that takes minutes, on the same database that runs the business.
Read Read: Slow ERP and back-office reports
Jump to
Audience
Type
48 of 48 articles
01
Start With a Problem
Recognize one of these?
Four situations we hear about most. Each one is a short read that names the cause and points to what to read next.
Slow product or catalog search
Searches that miss typos, filters with no counts, screens that time out past a million rows.
Read Read: Slow product or catalog searchA database that keeps growing, screens that keep slowing
Everything was fine at 200,000 rows. Adding indexes helped once. Then it didn't.
Read Read: A database that keeps growing, screens that keep slowingAI over private business data, without exposing it
Your team wants to ask ChatGPT or Claude about the numbers. IT wants the data to stay put.
Read Read: AI over private business data, without exposing it
02
Start With Your Role
Three reading paths.
ENGINEERING
What an index server is, how it differs from the database you already have, and how AI tools connect to it.
BUSINESS TEAMS
Why reports are slow, when to use a dashboard and when to just ask, and how to read the answer.
LEADERSHIP & IT
What leaves your environment, what doesn't, and how AI gets to private data safely.
Nothing matches. Clear filters.
03
DATA & INDEXING
The indexed data layer
Why a copy of your data, organized for finding and counting, makes application screens fast - and what it changes about your architecture.
PILLAR
What is an index server?
A copy of your data organized for finding and counting, not for storing and updating - and the reason the fast screens in large applications are fast.
COMPARISON
RDBMS vs index server: what's the difference?
They are not rivals. One is built for correct writes and relationships, the other for fast reads that search, filter and count. Most systems that grow end up needing both.
COMPARISON
Database index vs search index
"We already have indexes" is the first thing every engineer says. Both are called indexes. They answer different questions.
EXPLAINER
Why not query the production database directly?
Because it has a job already, and the reports compete with it.
EXPLAINER
What is faceted search?
Filters that come with counts. The user sees what is there before clicking, and the numbers come back in the same request as the results.
COMPARISON
Full-text search vs database LIKE
LIKE '%jaket%' matches characters. Full-text search matches words, forgives the typo, ranks the results and suggests the next one.
EXPLAINER
Incremental indexing: keeping an index fresh
An index is a copy, so it has to be refreshed. The question is how, how often, and what to tell users.
COMPARISON
OLTP vs OLAP: two kinds of database work
Transactions and analysis want opposite things from a database. The split explains most slow reporting screens, and where an index server fits.
COMPARISON
Cache vs read replica vs index
Three ways to take load off a database, and they are not interchangeable. Each one answers a different kind of read.
COMPARISON
Data warehouse vs index server
Both are copies of your data built for reading. One serves analysts with big questions; the other serves application screens with fast ones.
EXPLAINER
Apache Solr and LipiCore
Solr is the indexing foundation. LipiCore Engine is what your applications talk to. Here is what each layer does, and why you never call Solr directly.
EXPLAINER
When not to use an index server
An index server is a copy built for reading. Some screens need the original. Knowing which is most of the design.
WALKTHROUGH
Connecting your database: from tables to collections
Six decisions, in order: which screens, which collections, which fields, push or pull, what schedule, and which screen moves first.
04
DASHBOARDS & GRIDS
Shared numbers, working views
The right chart for the question, definitions everyone shares, and the saved views teams come back to every day.
COMPARISON
Dashboards vs chat: two kinds of questions
A dashboard answers the questions you ask every week with definitions everyone shares. Chat answers the question you have right now. Same data, same charts, different job.
EXPLAINER
Choosing the right chart for business data
Pick the chart by the question, not by taste. Change over time, comparison, share, distribution, relationship - each has one or two right answers.
COMPARISON
Dashboard vs report
A report is a point in time, sent to people. A dashboard is a live view, opened by people. Most teams need both and build the wrong one first.
EXPLAINER
KPI vs metric
Every KPI is a metric. Almost no metric is a KPI. The difference is whether someone will act when it moves.
EXPLAINER
Drilldowns and click to focus
A chart that answers "which region?" should let you click the region and ask the next question. Drilldown is how a dashboard holds a conversation.
EXPLAINER
Saved views as a workflow
Filter once, name it, come back to it every morning. A saved view is a question the dashboard keeps answering with current data.
EXPLAINER
Guided filtering
A filter that shows how many records each choice would leave is a filter you cannot get lost in. The counts do the guiding.
COMPARISON
Delivered analytics vs self-serve BI
Self-serve BI hands the team a tool and asks them to build. Delivered analytics hands them the dashboards, configured to the business, ready to use.
EXPLAINER
Access control: users, groups and scoped views
Not everyone needs to see everything. Access by user, permissions by group, and views scoped to a region or a team keep a shared dashboard relevant and safe.
EXPLAINER
White-label analytics
Your logo, your colours, your theme, on dashboards your teams and customers open every day. The vendor stays invisible unless you want it seen.
05
AI ON BUSINESS DATA
Exact answers in plain English
How an assistant turns a question into exact numbers, and what the model does and does not see.
PILLAR
How can AI answer questions about business data?
The model does not read your database. It turns your question into a structured query, runs it against indexed data, and writes up the result. The numbers come from the index; the words come from the model.
COMPARISON
Three ways AI answers from your data: RAG, text-to-SQL and indexed retrieval
"Is this RAG?" usually means "are the numbers right and is my data safe?" Three approaches, three different answers to that.
EXPLAINER
What is conversational BI?
Business intelligence you talk to. Ask in plain language, get a KPI, a chart or a table back, and ask the next question without building anything.
EXPLAINER
What the LLM sees
Not your database, not your index, not your collections. The question, the names it needs to map it, and the result it writes up. That is the list.
EXPLAINER
Bring your own LLM
The model is not part of the product. You choose the provider, hold the account and the key, and pay them directly. Change your mind later and nothing else changes.
EXPLAINER
How an answer gets its shape
Ask for one number and get a KPI tile. Ask for a ranking and get a leaderboard. The question decides the format, and the format decides what you see first.
EXPLAINER
Why grounded answers don't invent numbers
A model asked for a number will produce one. A grounded assistant produces the index's number, or says it cannot. The difference is where the figure comes from.
EXPLAINER
How to ask good questions of your data
Name the thing, the measure, the period and the breakdown, and an assistant can answer exactly. Leave them out and it has to guess, or ask.
06
MCP & AI TOOLS
Your data, in the AI tools your team uses
The open standard that lets ChatGPT, Claude and Cursor work with your data, and what a server for business data should look like.
PILLAR
What is MCP?
An open standard that lets AI apps connect to outside tools and data the same way, so a data source is integrated once instead of once per AI tool.
EXPLAINER
MCP server explained
A small service that publishes a list of tools and answers calls to them with structured results. For business data, the right tools are search, filter and aggregate over indexed collections - not raw SQL.
COMPARISON
MCP vs REST API: what's the difference?
A REST API is for programs written in advance. MCP is for AI clients that decide at runtime which tool to call. Most MCP servers sit on top of an API; it is not either/or.
COMPARISON
MCP vs RAG
RAG is a way of giving a model passages to read. MCP is a way of giving it tools to call. One is an approach to retrieval; the other is a plug.
EXPLAINER
The enterprise MCP server
One endpoint a whole company connects to, inside its own infrastructure, with keys it issues and a log it owns. What a team-grade MCP server needs that a laptop one does not.
COMPARISON
MCP server vs AI assistant: which one, or both?
The assistant is a ready-made conversation inside the product. The MCP server brings the same data into the AI tools your team already uses. Same answers, two doors.
07
SECURITY & DEPLOYMENT
Where things run, what crosses the boundary
Per product: where it runs, what leaves, to whom, and under whose control.
EXPLAINER
Data flow map: what leaves your environment, per product
One boundary, four flows. What goes in, what stays, what goes out, to whom, and who controls it - for each product separately, because the honest answer differs.
EXPLAINER
Runs in your infrastructure: what that means in practice
On your servers or in your private cloud, nothing hosted by the vendor, your keys. A phrase every vendor uses, and six things to check before believing it.
EXPLAINER
One-way data flow
Data goes from your systems into the index and never back. That one rule is why an indexed layer is easy to add, easy to audit and easy to remove.
EXPLAINER
Access keys and revocation
No key, no connection. Keys issued from your own deployment, one per team or tool, rotated on a schedule and revoked in a moment. The whole access model, in one page.
EXPLAINER
What counts as a record?
Pricing is by the records you index. A record is one row of a collection - one order, one product, one customer - so the count is a design decision, not a surprise.
08
START WITH A PROBLEM
Start from the symptom
Short reads that name the cause and point to the explainer.
PROBLEM
Slow ERP and back-office reports
The month-end report that takes minutes runs on the same database that runs the business. That is the problem, not the report.
PROBLEM
Slow product or catalog search
A search box backed by
LIKE '%term%' cannot forgive a typo, rank a result or count what is left after a filter.
PROBLEM
A database that keeps growing, screens that keep slowing
The screens were fine at 200,000 rows. At five million, the same queries crawl, and the next index will not save them.
PROBLEM
AI over private business data, without exposing it
The team wants to ask ChatGPT or Claude about the numbers. IT wants the data to stay where it is. Both are right.
PROBLEM
Every team has its own numbers
Sales and finance bring different totals to the same meeting, and the meeting is spent on which one is right. The numbers are not wrong. The definitions are.
PROBLEM
Reports that need engineering
Every new question becomes a ticket, and the queue is weeks long. By the time the report arrives, the question has changed.
Glossary
Thirty-six terms, one paragraph each. Index, facet, pivot, MCP, token and the rest.
Open the glossary
Live Demo
See it on real data.
The demo instance runs dashboards, data grids and the AI Assistant on real business data. No sign-up.