There is an easy way to misunderstand Opis App Hub.
Open the navigation, see a long list of research tools, and assume you are looking at 22 separate applications that all need to be learned individually.
You are not.
The names change because the business question changes. A Supplier Finder asks for different information from a SWOT analysis. A Competitor Finder has a different job from an Authority Auditor. A Writer's Brief needs different inputs from a Launch Readiness assessment.
But underneath those different research modes sits the same operating system.
That distinction matters because it changes the onboarding problem completely. You do not need to memorise 22 interfaces. You need to understand one repeatable way of working:
Choose a question → provide context → research → inspect the result → challenge what is uncertain → refine → reuse or deliver.
Once that rhythm becomes familiar, moving from one research mode to another should feel less like opening a new piece of software and more like asking the same research team to investigate a different problem.
This guide explains that operating model.
Choose
a business question
Research
build the evidence base
Interact
clarify and improve
Start with the three-panel mental model
The easiest way to orient yourself inside Opis is to think in three panels.
Panel 1 — Navigation
This answers a simple question: Where do I want to go?
Use it to move between research modes, workflows, saved work, Content Studio and the wider workspace.
You should not need to understand the engineering behind a research mode before selecting it. The title tells you the business job. If you want to identify potential suppliers, choose Supplier Finder. If you want to understand a competitor's website, choose Website Competitor Analysis. If you want a PESTEL assessment, choose PESTEL.
Panel 1 is about choosing the job, not operating the engine.
Panel 2 — Results & Workspace
This answers: What did Opis find?
It is the main working area. Your research result, structured findings, scores where applicable, evidence, timelines, charts and report content live here.
Treat this panel as a workspace rather than a static answer page.
The first result is not something you are expected to accept blindly. It is the current state of the research: what has been found, what can be supported, what remains uncertain and what deserves your attention next.
Panel 3 — Interaction
This answers the most important question: What do I want Opis to do with what I now know?
This is where the relationship stops being one-directional.
You can provide additional context, correct an assumption, give an instruction or ask a question about the current research. Opis keeps a refinement trail so that the result can evolve without pretending the first pass was perfect.
The mental model is therefore very simple:
Navigate → Understand → Interact.
That pattern is more useful to learn than the individual names in the application catalogue.
What actually changes between research modes?
Far less than you might think.
Opis uses a configuration-driven system. Each research mode defines the things that are specific to that job: the questions it needs to ask, the evidence it should seek, the structure of the output and, where relevant, the rules used for deterministic scoring.
The surrounding machinery is shared.
That shared machinery can include grounded search, first-hand website reading, official statistics, source handling, evidence checks, the input ledger, assumptions and gaps, ICC refinement, exports and downstream reuse.
This is why a relatively small configuration can produce a serious research workflow. The value is not hiding in a giant prompt attached to every menu item. The value comes from connecting a particular business question to a common evidence and output system.
For the user, that is good news. Familiarity compounds.
Learn how to inspect sources once and that skill transfers. Learn how to refine an ambiguity once and that skill transfers. Learn how to open a finished report in your Library or pass research into Content Studio and that skill transfers too.
Your first run: what to do, in order
For a first session, resist the temptation to explore everything.
Choose one research problem you genuinely understand. Competitor Finder is a good example because you will usually know enough about your own market to spot whether the result makes sense.
Step 1 — Give Opis a clear objective
Do not try to write an enormous prompt.
Tell the system what you are investigating and why.
For example:
“We are a Leeds-based B2B coffee supplier considering expansion into office subscriptions. I want to understand which local competitors are strongest in corporate delivery and where their proposition appears weak.”
That is far more useful than simply writing “find my competitors.”
You have provided geography, business model, customer type and a decision objective.
Step 2 — Use Tailor when the objective needs more context
Opis can use its adaptive intake process to identify a small number of additional optional questions that are material to the job.
The important word is optional.
The aim is not to make you complete a consultancy questionnaire before you are allowed to research anything. It is to discover whether one or two missing pieces of context could substantially improve the result.
If you know the answer, provide it. If you do not, continue.
Step 3 — Let the first research pass establish the evidence base
Opis combines the supplied brief with the evidence behaviours configured for that research mode.
Depending on the job, that can include search grounding, reading selected websites directly, consulting official statistical sources and carrying verified evidence from an earlier workflow stage.
The output is not just prose. It is stored as structured research with provenance, confidence information and the wider reliability envelope that allows the same intelligence to be used elsewhere in the system.
Step 4 — Read the uncertainty before you read the conclusion
This is one of the best habits you can build in Opis.
Before becoming attached to the headline answer, inspect:
- assumptions;
- ambiguities;
- contradictions;
- gaps;
- sources;
- confidence.
Those sections tell you where the boundaries of the result are.
If a competitor's pricing could not be confirmed, that is operational information. If two sources disagree about a service area, that matters. If the system had to infer whether a company serves B2B or B2C customers, you should know before you make a decision based on the output.
Step 5 — Interact with the result
Now use Panel 3.
Perhaps you know something the public web does not:
“They stopped serving consumers last year. Their current business is almost entirely corporate contracts.”
Add it as context.
Perhaps you want the analysis to behave differently:
“Prioritise recurring contract value over one-off order pricing.”
Add it as an instruction.
Perhaps you want to interrogate the current evidence:
“Do the existing sources actually establish that they offer next-day delivery?”
Ask the question.
The point is not to become a better prompt engineer. The point is to turn the first research pass into a dialogue between public evidence and your real-world knowledge.
Step 6 — Reuse the finished intelligence
A completed result does not have to die on the page where it was generated.
It can become a report, be saved and revisited, contribute to a collected profile, move into a workflow, provide research context to Content Studio, be transcreated for another language or be delivered through one of the supported integration paths.
That is why thinking of the system as “22 apps” can be misleading.
The menu gives you 22 starting points. The useful thing is what happens after the research begins.
A note about scores
Not every Opis research mode needs a score.
Where deterministic scoring is configured, Opis deliberately separates the job of sensing information from the job of calculating the evaluation. The research layer gathers the relevant attributes and evidence; the scoring layer applies fixed rules.
This matters because a language model is very good at interpreting language but is not the right place to hide a supposedly stable business scoring methodology.
If new authoritative context changes a scored input during refinement, the deterministic score can be recomputed from the revised information.
The practical lesson for a new user is simple:
Read a score as the output of declared criteria, not as the AI's mood.
Your first-hour onboarding exercise
If you are using Opis for the first time, try this rather than randomly opening every feature.
- Choose one company or product you know well.
- Run one appropriate research mode.
- Open at least three of its sources.
- Read every assumption and gap.
- Correct one item using the Interaction Panel.
- Ask one question the existing evidence may or may not be able to answer.
- Save the result.
- Open it again from your Library.
You will have learned more about how Opis works than you would by clicking through all 22 menu items.
What to remember
Opis is designed so that complexity lives underneath a consistent operating model.
You choose the business question in Panel 1.
You work with the research in Panel 2.
You challenge, clarify and direct it through Panel 3.
Everything else builds on that.
Learn that once and the rest of the platform becomes much easier to understand.
Next in the series: why the first AI answer should be treated as the beginning of research, not the end.
Learn the shared workspace before trying to memorise every research mode. Once navigation, results and interaction feel familiar, each new mode becomes a different business brief inside a system you already understand.
Frequently Asked Questions
No. The fastest route is to learn the shared workspace first, then choose research modes as different starting points for specific business questions.
The working model is Navigation, Results & Workspace, and Interaction. The Interaction panel is where questions, clarifications and additional context are handled.
Start with one real business question, inspect the evidence and uncertainty signals, add one useful clarification, and learn where the finished intelligence can be reused.