Blog
What we have worked out so far.
How website assistants actually work: what to train them on, how retrieval behaves, when to hand a conversation to a person, and how to read the analytics. No case studies yet.
Where to start
Four we would hand someone first, rather than the four that happen to be newest.
Retrieval augmented generation, a practical explanation
For a website assistant, retrieval quality dominates model quality. Here is why, and what that means for where you spend effort.Vishal Chiniwar2026-04-01When to hand off to a human, and how to decide
Handoff triggers should be written rules you can audit, not a judgement made fresh each time.Sachin Aathreyaa K M2026-04-29Choosing what to train a website assistant on
More sources produce worse answers, not better ones. Source selection is the main quality lever you control.Sachin Aathreyaa K M2026-06-17
Showing 12 of 28 articles
Adding an assistant to a Webflow site
Where the embed goes in Webflow, what breaks, and the constraints specific to the platform.Vishal Chiniwar2026-06-03Latency budgets for a chat widget
Perceived latency is a UI problem as much as a model problem. Here is where the time goes and which parts are worth optimising.Vishal Chiniwar2026-02-18What to tell visitors about AI chat and data
The disclosure that belongs next to the chat box, and how to write it without turning it into a legal document.Vishal Chiniwar2026-01-28Measuring content gaps without a research budget
You do not need a user research programme to find out what your website fails to explain. The people already visiting will tell you for free.Sachin Aathreyaa K M2026-07-08Adding an assistant to a WordPress site
Three ways to add the embed, which one to choose, and the caching and consent issues specific to WordPress.Vishal Chiniwar2026-05-27One click deploy: what actually has to be true first
Embedding a widget takes a minute. Making it worth embedding takes longer. Here is the honest checklist.Vishal Chiniwar2026-06-10What a good handoff actually contains
Escalating a conversation without context makes the visitor start over. Here is the minimum payload a human needs.Sachin Aathreyaa K M2026-04-22Questions to ask any AI chat vendor before you sign
A checklist for evaluating a website assistant, including the questions we would have to answer ourselves.Sachin Aathreyaa K M2025-12-17Chat analytics that change decisions
Message volume is a vanity metric. Here are the four numbers that actually cause someone to do something differently.Vishal Chiniwar2026-04-08Qualifying without an interrogation
Two questions asked at the right moment beat six asked upfront. Most qualification signal is already in the transcript.Sachin Aathreyaa K M2026-04-15Where your website content goes when you train an assistant
Draw the data flow before you promise anything about it. Here is the map, and the questions it forces you to answer.Vishal Chiniwar2026-02-11Widget placement, and when not to show a chat launcher
A chat launcher on every page at all times is a default, not a decision. Restraint improves both conversion and how the brand reads.Sachin Aathreyaa K M2026-05-13Reliability when a model provider goes down
Provider incidents happen. What your assistant does during one is a design decision you make now or by accident later.Vishal Chiniwar2026-01-14Rate limits, abuse and cost control
A public endpoint that calls a paid API is a spending endpoint. Budget guards belong in the design, not in an incident review.Vishal Chiniwar2026-01-07Chunking website content for retrieval
Chunk boundaries decide answer quality more than chunk size does. Where you cut matters more than how big the pieces are.Vishal Chiniwar2026-03-25Retention decisions are product decisions
How long you keep conversation data determines what analysis is possible. Copying someone else's number decides your product by accident.Vishal Chiniwar2026-01-21Lead capture inside a conversation without killing it
Asking for an email before answering anything is the fastest way to end a conversation. Sequence matters more than form design.Sachin Aathreyaa K M2026-05-06Support tickets and pre sale questions are different signals
Mixing post sale support with pre sale questions produces an average that describes nobody. Separate them and both become useful.Sachin Aathreyaa K M2026-07-15Reading intent from question phrasing
How a visitor phrases a question predicts what stage they are at more reliably than the page they were on when they asked it.Sachin Aathreyaa K M2026-06-24Data boundaries, a plain explanation for site owners
What you can and cannot honestly tell your visitors about an AI assistant on your site, without a legal team.Vishal Chiniwar2026-02-04The questions your pricing page fails to answer
Most pricing objections are not about price. They are about missing information that the pricing page could have supplied.Sachin Aathreyaa K M2026-07-22Why FAQ pages go stale and what replaces them
An FAQ page written once at launch describes the questions a team imagined, not the questions people ask. Here is how to keep one honest.Sachin Aathreyaa K M2026-07-01Keeping an assistant in sync with a changing site
Retraining cadence is a product decision, not a maintenance chore. Here is how to decide it rather than default to it.Vishal Chiniwar2026-05-20Handling questions the assistant should refuse
Refusal design is a product feature. Deciding the boundary in advance is what stops confident wrong answers.Sachin Aathreyaa K M2026-02-25Embeddings, what they do and what they do not do
Embeddings measure similarity, not truth and not relevance. Understanding the difference prevents a specific class of bad answer.Vishal Chiniwar2026-03-18Measuring answer quality without a labelled dataset
You will not have a golden dataset. Two proxies get you most of the way, and both are already being generated.Vishal Chiniwar2026-03-04Why multi model routing exists
Cost, latency and task fit differ by query. Routing is how you stop paying frontier prices for questions that do not need them.Vishal Chiniwar2026-03-11How to turn chat transcripts into a content roadmap
A repeatable monthly process for converting raw chat logs into a prioritised list of page briefs, without a research budget or a dedicated analyst.Vishal Chiniwar2026-07-29
Nothing matches that. or try the glossary.
About the blog
Sachin Aathreyaa K M, Co-founder, CEO and CPO at Creoglyph. They are written alongside building Creobot rather than as a content programme, which is why they cover decisions rather than trends.
When something is worth writing down. There is no publishing schedule, and an article gets revised when the underlying decision changes rather than being left as a dated snapshot.
Mostly not. They describe how website assistants behave in general. Where Creobot makes a particular choice, the feature pages say so rather than the article.
The four under Where to start, above. If you want the mechanism, read the retrieval explainer. If you want the commercial argument, read what visitor questions reveal.
No. Creobot has no published customer stories yet, so there is nothing to write a case study about. Any figures in an article are illustrative and labelled.
Yes. The search box matches title, description and tags, and the category chips filter the grid. Both update the count and share one empty state.
Launch
Building something similar?
Creobot opens to teams in order. Access opens to teams in order and access is by waitlist. Ask a question and we will answer it directly.