CONNECT · PROJECT CONTEXT
Give your AI tools the business context they are missing.
Keep a shared, editable brief for each website: what the business does, its audience, goals, competitors and key pages. Every connected agent works from the same context instead of reconstructing it in each conversation.
- One brief per project
- Competitors and key pages
- Read by every agent
- BusinessPhone and accessories retailerNairobi, ships across Kenya
- GoalGrow non-branded salesPhones under KSh 30,000
- Competitorscompetitor-a.co.ke, competitor-c.co.keWith notes
- Key pages/phones/ · /accessories/Role and target topic
Brief
business, audience, positioning and goal
Competitors
names, domains and notes
Key pages
role and target topic for each
Research log
paid research agents ran
INSIDE THE PRODUCT
One brief. Every tool. Every conversation.
Write the context once and let research, reports and agents use it.
The brief
Explain the business once
Record what the business does, who the audience is, the site’s positioning and the current objective, plus custom sections for anything else your team needs.
- Business, audience and positioning
- Current objective
- Custom sections
| Field | Value | Updated |
|---|---|---|
| Business | Phone retailer | 12 Oct |
| Audience | Price-conscious buyers | 12 Oct |
| Objective | Non-branded growth | 20 Oct |
| Tone | Plain, practical | 12 Oct |
Competitors & pages
Keep competitors and key pages in one place
Maintain the competitors with their domains and notes, and the key pages with their role and target topic. Competitor tracking and AI visibility use the same list.
- Competitors with notes
- Key pages with role and topic
- Shared with competitor and AI tools
- Money/phones/Target: buy phones in Kenya
- Money/phones/redmi-note-14/Target: redmi note 14 price
- Guide/blog/best-phones-2026/Target: best phones 2026
- Trust/delivery/Target: phone delivery Nairobi
Research log
See the research agents already ran
Agents record the paid research they run in a research log, so the next conversation builds on it instead of paying for the same lookup again.
- Paid research recorded
- Visible to the whole team
- Avoid repeat lookups
| Research | By | Date |
|---|---|---|
| Keyword ideas: phones | Claude | 27 Oct |
| Competitor gap: competitor-c | Codex | 24 Oct |
| SERP: redmi note 14 | Claude | 21 Oct |
Research log
Reuse paid research instead of repeating it
Paid research is logged against the project, so the next agent session can read what was already looked up before spending credits again.
- Paid research logged
- Reused by the next session
- Fewer repeat requests
- Keywordsbudget phones, Kenya24 Oct
- Competitorscompetitor-a gap22 Oct
- SERPphone a price20 Oct
- AI24 prompts checked14 Oct
Website setup
Let an agent draft the first brief
An agent can research the project website and propose the business description, audience and competitors for you to review and save.
- Researched from the website
- Suggested competitors to review
- You review before saving
| Field | Suggestion | Status |
|---|---|---|
| Business | Phone and accessories retailer | Review |
| Audience | Buyers in Nairobi and Mombasa | Review |
| Competitors | 3 domains | Review |
| Key pages | 6 URLs | Review |
Competitors
One competitor list for every module
The competitors saved in project context are the ones competitor analysis compares, rank tracking shows beside your positions and AI visibility looks for in answers, so every view argues from the same set.
- Shared by research and tracking
- Suggested, then reviewed
- Updated in one place
| Domain | Used in | Added |
|---|---|---|
| competitor-a.co.ke | Gaps · ranks · AI | 12 Oct |
| competitor-b.co.ke | Gaps · ranks · AI | 12 Oct |
| competitor-c.co.ke | Ranks · AI | 20 Oct |
WHAT IT MEASURES
What project context holds.
What each metric tells you, and where it comes from.
| Metric | What it tells you | Source |
|---|---|---|
| Business | What the business does and sells. | You |
| Audience | Who the website serves. | You |
| Goals and positioning | The current objective and how the business wants to be seen. | You |
| Competitors | The competitive set used across research. | You, with suggestions |
| Key pages | The pages that matter most. | You |
| Writing preferences | Tone and style for reports. | You |
UNDER THE HOOD
How project context is used.
How the numbers are produced, so you can judge them for yourself.
Stored per project
The business, audience, goals, positioning, competitors, key pages and writing preferences are stored for each project.
Read first by agents
Connected agents read the context at the start of a task, so research and reports reflect the business without a long prompt.
A log of paid research
Paid research is logged with the project, so later sessions can reuse it instead of paying again.
Drafted, then reviewed
An agent can research the website and suggest a first draft; you review and edit it like any shared document.
HOW IT WORKS
A brief every agent can share.
Four steps from connecting your data to a decision you can stand behind.
Draft it
Write it yourself, or let an agent research the website and suggest a draft.
Review it
Check every field; nothing is saved without you.
Use it
Connected agents read it before research and reports.
Keep it current
Update it when goals or competitors change.
WITH YOUR AI AGENT
Ask your agent to start from the brief.
Agents read the shared context first, so they stop asking the same questions.
- Uses the same data as the web app
- Paid tools estimate their cost first
- Saves the report to the project
YouBefore you start, what do you know about this project?
get_project_context
Agentyourstore.co.ke sells phones and accessories to buyers in Nairobi and Mombasa. The goal this quarter is non-branded growth. Competitors: competitor-a, competitor-b and competitor-c. Six key pages are listed.
YouAdd competitor-d to the competitors.
AgentAdded competitor-d to the project’s competitors. Every agent and report now uses the updated set.
Illustration · sample data
USE CASES
Built for the way you work.
How teams put it to work.
Brief agents per client
Each client project carries its own brief, so agents never mix up businesses.
Keep everyone aligned
The goals, competitors and key pages live in one place that people and agents both read.
Onboard a new client once
Write the client brief once and every agent session starts from it.
DATA & LIMITS
Know what the numbers can and cannot tell you.
Where the data comes from, how fresh it is, and the questions it cannot answer on its own.
| Source | What it provides | Freshness & limits |
|---|---|---|
| You and your team | The brief itself. | Only as current as the last update. |
| Website research | Suggestions for a first draft. | Reviewed before saving. |
A brief, not a vault
Do not store passwords, keys or private customer records in the context.
Agents follow it
Connected agents and the optional in-app Sam beta read it; only connect agents you trust.
It can go stale
Update it when priorities change, or agents will work from old goals.
Suggestions need review
Website research proposes; you decide.
WORKS WITH
Connected to the tools you already use.
The sources and clients this feature reads from or works alongside.
CHOOSE THE RIGHT VIEW
Which view answers which question.
Start where the question points. Each view keeps its own source.
| Your question | Start with | Why |
|---|---|---|
| What should agents know about us? | Project context | The shared brief. |
| How should reports look? | Reports | Templates and saved reports. |
| Who are our real competitors? | Competitor analysis | Gaps across the set. |
LEARN & TRY
Guides and free tools for the same job.
Read how the evidence works, or try a free check before you connect a project.
Content brief builder
Draft a brief from a target query.
Your data vs market estimates
Which question first-party data and provider estimates each answer.
FAQ
Project context: common questions.
Straight answers about what it does, what it costs and where its limits are.
Who reads the project context?
Connected AI agents and the optional in-app Sam assistant, through the project’s MCP tools.
Can I store passwords or keys here?
No. Project context is a working brief. Do not store credentials or private customer records in it.
Where do competitors come from?
You add them, or review suggestions. The same list is used by competitor tracking and AI visibility.
Does the research log cost anything?
No. It records research agents already ran, so it can be reused.
Can an agent change the context?
Yes. Agents with access to the project can update it through MCP, so connect only agents you trust and review what changed.
Is the context shared across projects?
No. Each project has its own context.
What is the research log?
A record of paid research done for the project, so later sessions can reuse it.
What should I not put in the context?
Passwords, API keys and private customer records. The context is a working brief that agents read.
Does Sam use the context?
Yes. The optional in-app Sam beta reads the same context as connected agents.
What are the limits of the data?
Project context is read by connected agents and the optional in-app Sam beta. Only connect agents you are authorised to use, and do not store secrets in the context.
How do I get access?
Platform access is reviewed. Request access with your website and the data sources you use, and the team will help you connect them.
See it with your own data.
Tell us about your website. We will review access and help you connect the right sources.