A
Risk inbox
A ranked queue of what is slipping, stuck or newly at risk — evidence attached.
AI operations radar Internal + external signals Pre-seed
The operations radar for fast-moving companies.
Companies run on a dozen tools and nobody sees the whole picture until something breaks. Retenna connects those tools — and the outside world that moves them — into one live view, then flags what is slipping, stuck or newly at risk while you can still act. You decide what it handles on its own and what it runs past you first.
Retenna · company-context radarLive view
Illustrative — signals shown are examples, not live customer data.
Connects
Your whole stack — chat, work trackers, code, CRM, billing, the calendar — wired to the outside world that moves it: customers, competitors, new rules.
Catches
Work slipping, commitments quietly at risk, decisions stuck for weeks, a competitor's move that just put a deadline in doubt. Retenna spots it while there's still time to change the outcome.
Acts
Every flag arrives with the evidence, the next step and the right person to take it. Let the routine fixes run on their own; keep the judgement calls for yourself. You set the line.
Pre-seed. Figures marked Placeholder or Assumption are modelled and not yet validated. No live customer traction is implied.
01Problem
Companies have more tools than ever, but less operational clarity.
The system of record for work is spread across a dozen tools. The view of the whole company — what is on track, what is slipping, who is on what — is still pieced together by hand.
CONTEXT · FRAGMENTEDManual
Illustrative — the only link between sources is manual effort: thin and incomplete.
Where context lives today
- Work lives in Linear, Jira and Asana.
- Decisions live in Slack and Teams.
- Strategy lives in Notion, Confluence and Drive.
- Customer commitments live in Salesforce, HubSpot, Zendesk and Intercom.
- Delivery risk lives in GitHub and CI.
- Time pressure lives in calendars and the on-call rota.
- External change — competitors, the market, new rules — lives outside the company entirely.
The gap
Each tool answers its own slice. The company-wide question — are we okay, and what needs me today? — has no home. So it gets answered in meetings, from memory, and usually a step too late.
The problem, in one lineOff-track
The system of record for work is fragmented. The operating view of the company is still manual.
02Why now
This was not buildable three years ago.
Reading a whole company at once needs a model that can hold messy context, a reason the old way stopped working, and a head start before the agents arrive. All three landed in the last couple of years. They are still landing.
CAPABILITY · TIMELINEInflection
Illustrative trajectory — the curve is schematic, not measured data.
Enabling shifts06 · converging
- S-01Models reason over messy context
Until recently a model could answer one clean question. Frontier models now read across a dozen scattered sources at once and hold a coherent picture.
- S-02More to do, fewer people
Teams are leaner and carry more. The work of keeping everyone aligned used to grow with headcount; now there is no headcount to grow it with.
- S-03Distributed work scatters the signal
Remote and async work spreads decisions and status across tools and time zones. Nobody is in one room watching it any more.
- S-04The data is already there
Companies already pour the signals into their SaaS tools every day. Nothing reads them together. The raw material is sitting unused.
- S-05Agents need somewhere to stand
An agent cannot do anything useful in a business it does not understand. It needs grounded company context first, and that has to be built before the agents arrive.
- S-06From summarising to operating
AI has spent two years answering questions about the past. It is starting to act on the present. That turn is happening now.
- The first wave of AI looked backwards: search and summarise what already happened.
- The next one looks at the present, decides what matters today, and does something about it.
03Solution
Retenna turns company context into an operating loop.
Connect the tools, build a living context graph, watch the signals, detect drift, explain the evidence, recommend the action — then track whether it resolved. Every cycle teaches the graph which risks were real.
Operating loopContinuous
Illustrative — the loop runs continuously; stages shown are the operating model, not live customer data.
- 01Connect tools
Plug into the systems a company already runs.
- 02Build context graph
Link goals, work, people, customers — and outside events.
- 03Watch signals
Read internal activity and external change.
- 04Detect drift
Spot what is off-track, blocked or unassigned.
- 05Explain evidence
Trace the signals behind every flagged risk.
- 06Recommend action
Name the next move and the right person for it.
- 07Track resolution
Confirm whether it actually resolved.
The loop is closed: stage 07 feeds back into the context graph, so resolved risks and effective actions are recognised and the next cycle is sharper.
04Product
Eight modules, reading from one picture.
Not a sprawling suite — a handful of surfaces that all draw on the same company context, so a billing slip, a quiet customer and a competitor’s launch show up in the same place, already connected.
RETENNA · RISK INBOXLive
Flag · INS-2041
Off-trackLaunch will slip
The billing redesign is booked to ship Friday. It is going to miss — and so far nobody has said so.
Why Retenna thinks so4 signals
- LinearFour blocking issues are still open.
- GitHubNo pull request has merged in five days.
- CalendarBoth engineers on it are away next week.
- HubSpotLaunch comms are already booked for Friday.
Next step
Move the launch comms, pull two issues forward, and put the go/no-go on Wednesday’s agenda — routed to the engineering lead.
Set this kind of fix to run on its own, or keep it approve-first — you choose, per workflow.
Illustrative flag — example signals, not live customer data.
B
Monday brief
The weekly read on what changed, what is blocked, and what needs a call.
C
Context graph
Goals, projects, people, customers and commitments, linked across every tool.
D
Decision queue
The open calls that have stalled — surfaced before they block the work.
E
Signal monitor
Market, competitor, regulation and news mapped to the work they actually touch.
F
Actions
Drafted for you to send — or set to run on their own. You choose, per workflow.
G
Evidence trail
The signals behind every flag, so nothing is a black box.
H
Outcome tracking
Whether the risk cleared, and whether the action actually worked.
Also on the radar this week
Account going quiet
At riskA top customer filed three urgent tickets, last month's invoice is 20 days overdue, and the shared channel has gone silent — six weeks before renewal.
ZendeskStripeSlackSalesforce
A competitor moved
Newly importantA rival shipped a feature three open deals had asked for. Retenna tied the news to the roadmap epic and those deals, and asked whether to pull it forward.
NewsJiraSalesforce
Burn vs. pipeline
WatchCloud spend jumped this week while two enterprise deals stalled. Retenna connected the spend to the runway and flagged the gap before the board call.
DatadogSalesforceQuickBooks
The point is the joins. No single tool sees a quiet Slack channel, an overdue Stripe invoice and a looming renewal as one story. Retenna does — and it connects a competitor’s launch to next quarter’s roadmap the same way.
05Competition
Every tool owns a lane. Retenna reads across all of them.
Teams already run a stack — Linear and Jira for work, Slack for talk, Notion for docs, Salesforce and HubSpot for customers, Looker for numbers, Copilot to draft. Each is good at its job. None of them connects the lanes, watches the outside world, or tells a leader what to do next.
Capability matrix · the stack vs. the radarScored honestly
| Tool / category | Stores the worksystem of record | Searches knowledgeretrieval | Summarisessynthesis | Tracks goals & projectsstatus | Links internal + externalacross tools | Spots what's slippingdrift | Names the next stepwho + what | Tracks the outcomedid it get fixed | Built for how leaders run the weekfor leaders |
|---|---|---|---|---|---|---|---|---|---|
| Linear · Jira · Asana · ClickUpWork tracking | |||||||||
| Slack · Microsoft TeamsMessaging | |||||||||
| Notion · Confluence · Google DriveDocs & knowledge | |||||||||
| Salesforce · HubSpotCRM & RevOps | |||||||||
| Glean · enterprise searchSearch layer | |||||||||
| Copilot · ChatGPT Enterprise · GeminiGeneral assistants | |||||||||
| Looker · Amplitude · PostHogBI & product analytics | |||||||||
| RetennaOperations radar |
FullPartialNone
Illustrative. Tool names are plain text — each category is strong in its own lane.
Where we sit
We are not trying to replace the stack. Retenna sits across it — the layer that reads every lane, watches the outside world, and turns it into the next move.
06Defensibility
Anyone can summarise. The moat is the context behind it.
An incumbent can copy a summary in a weekend. What it cannot copy is a labelled, cross-tool history of how one company actually runs — and that only builds with use.
Moat · layersCompounding
- OUTERSurface
- L1Connected
- L2Lived
- CORECompounding
The outer rings are easy to imitate. The core is earned, one company at a time.
Moat · registerL0 → L3
- CopyableSummariesL0
Any model can compress a thread. Table stakes, not a moat.
- Builds with dataCross-tool context graphL1
Goals, projects, customers and commitments, linked across every tool.
- Builds with dataInternal + external signal mappingL1
What the company is doing, read against what the market is doing to it.
- Hard to copyHow the company runsL2
Who decides what, how things escalate, and whether they actually get done.
- Hard to copyLabelled outcomesL2
Which flags were real and which fixes worked, graded as the weeks pass.
- Hard to copyWorkflow embeddingL2
Built into the leadership meeting and the Monday brief, week after week.
- Hard to copyPermission-aware contextL2
One picture that still respects who is allowed to see what.
- Compounds with usePlaybook libraryL3
The drift patterns that recur and the responses that fixed them, reusable.
- Compounds with useCustomer-specific baselinesL3
What normal looks like here, modelled per account so abnormal stands out.
- Compounds with useTrust & security architectureL3
The access and governance posture that earns the right to this data.
- Compounds with useGoverned agentsL3
In time, agents that already know the company's priorities and policies.
Each week deployed adds labelled history a fresh copy starts without.ASSUMPTION
07Market & model
A bottom-up path to $100M ARR.
We size this from the companies we can actually win, not a top-down TAM slide. Land cheap with the audit, grow the account as Retenna connects and watches more of the business. Every figure below is a labelled assumption.
MODEL · BOTTOM-UP (ASSUMPTIONS)Modelled
TARGET$100M ARR
TARGETREAD BOTTOM-UP · thousands of startup, mid-market and enterprise accounts compound the three bands toward the target. Each band is an independent assumption.
Customer counts and ACV bands are illustrative assumptions, not market research or booked revenue.
PRICING · INDICATIVELands on Growth
- Ops risk auditThe way in — a first read on a company's operating risk.Free / paid
- StartupSmall teams keeping delivery on track without a chief of staff.$500–$1,500/ mo
- GrowthTypical landingScaling companies where drift outruns manual oversight.$1,500–$5,000/ mo
- ScaleOrganisations standardising across many teams, with controls.Annual
PLACEHOLDER
Expansion · later
Usage-based pricing follows the value: the more of the business Retenna connects and watches, the more the account is worth.
Connected toolsPeople watchedSignal volumeAgents & actionsHistory depth
Categories we touch
Work managementEnterprise searchKnowledge managementAI productivityBusiness operationsRevOps / CSOps / ProductOpsAI agents for internal ops
The whole model is an explicit assumption set, not a market-research claim. No sourced TAM, no booked revenue and no traction implied.
08Go-to-market
Land with the Monday brief. Grow into the operating layer.
One sharp workflow gets us in the door; the context we build to deliver it becomes the platform. A motion the founder can run today — and one that produces the case studies the seed round needs.
ChannelsFounder-led outboundInstantly.ai emailLinkedIn contentFree ops risk auditWhite-glove setup
GTM · MOTIONFounder-runnable
- 01Cold outbound
- 02Book the audit
- 03Connect tools
- 04First Monday brief
- 05Paid pilot
- 06Subscription
- Reach the ICP
- Free / paid
- A handful
- Real risks, fast
- Convert
- Monthly / annual
Illustrative — stages and conversion are modelled, not yet measured.
WEDGE · WHERE WE LAND
- B2B SaaS and startups, 15–150 people
- Running Slack, Linear or Jira, GitHub, Notion or Drive
- A founder or COO still chasing updates by hand
What changed, what is blocked, what is slipping, what needs a decision — and the right person for the next step.
Then it grows
Weekly briefDaily risk inboxAction routingGoverned agents
WHO BUYS
- Founder / CEO
- COO
- Chief of Staff
- Head of Ops
- Head of Product
- Engineering lead
Sell to the person who spends Monday morning chasing updates. The audit pays for itself the first time it catches a slip they would have found on Friday.
The people who feel the status problem directly.
ASSUMPTIONConversion through each stage is modelled, not yet measured — early outbound is designed to produce the first validated case studies.
09Roadmap
From a manual weekly brief to a governed operations radar.
Three horizons. Each one has to earn the next: ship the brief, then connect natively, then let the radar run live. Here is what is built versus what is planned.
ROADMAP · HORIZONSH1 in build
H1 · NowMVP
In buildEarn the first weekly brief from whatever a team uploads or connects lightly.
- Manual context onboarding
- CSV / document upload
- Slack export / OAuth
- Linear / Jira — basic integration
- GitHub — basic integration
- Weekly operations brief
- Risk inbox
- AI evidence explanations
- Draft actions
H2 · NextMLP
NextConnect natively, watch more of the business, and set who can act on what.
- Native Slack / Teams
- Native Linear / Jira / GitHub
- Notion / Google Drive
- Calendar
- HubSpot / Intercom
- External watchlists
- Decision queue
- Action approval
- Audit logs
- Role-based access control
H3 · VisionFull product
VisionA live radar over the whole company, with agents that act inside the limits you set.
- Company context graph
- Real-time operations radar
- Governed AI agents
- Customer-commitment tracking
- Finance / burn / vendor risk
- Product / engineering delivery risk
- Market / competitor / regulation monitoring
- Operating-rhythm automation
- Board / leadership brief generation
- Enterprise permissions and compliance
H1 is in build. Nothing on this rail is live yet.
TARGETDates are a target we are working towards, not a promise. Nothing here ships before the MVP does.
10Risks
What could sink this, and what we are doing about it.
Eight risks we take seriously, each paired with the move we are already making on it.
RISK · LEDGER8 named
The riskWhat we are doing
- R-01RiskProduct too broadChasing every job at once dilutes the wedge.MitigationShip one thing first: the weekly ops brief.
- R-02RiskIntegration complexityEvery connector is surface area to maintain.MitigationShip a limited set of integrations first; depth before breadth.
- R-03RiskData qualityGarbage in undermines every recommendation.MitigationEvidence-backed recommendations only; always show the trail.
- R-04RiskAI hallucinationA confident wrong answer erodes trust fast.MitigationEvery claim cites its evidence; you choose what auto-runs and what waits for a yes.
- R-05RiskSecurity & privacyWe read sensitive company context.MitigationRead only what we are granted; respect each tool's permissions.
- R-06RiskIncumbentsEstablished tools can bolt on features.MitigationBecome the place the company reads each week; build a track record of outcomes they can't.
- R-07RiskHard ROIOperational value can be diffuse.MitigationMeasure time saved, risks caught, blockers resolved, commitments protected.
- R-08RiskBuyer confusionA vague pitch wins no one.MitigationOne ICP, one workflow, one sentence anyone can repeat.
Each risk sits next to the mitigation already under way.
This is the list as it stands today. A mitigation is a plan, not a promise. We will update the ledger as we learn what holds.
11Validation
No traction yet. Here is the bar we have set and the work behind it.
Pre-seed and pre-revenue. The live metrics below stay empty until customers fill them. The numbers on the right are targets we have set ourselves, not results.
VALIDATION · READOUTS (EMPTY)Pre-revenue
[ — ]
Companies analysed
Cumulative
[ — ]
Risks surfaced
Detected
[ — ]
Actions resolved
Closed
[ — ]
Hours saved
Modelled
[ — ]
Discovery calls
Completed
[ — ]
Design partners
Signed
[ — ]
MRR
Recognised
If a number is not real, we leave it blank. These will fill as we sign customers.
No live customer data yet. Nothing to dress up.
Targets before seed
The bar- 20Discovery calls completedTARGET
- 5Design partners signedTARGET
- 2Paid pilots runningTARGET
- 1–2Quantified case studiesTARGET
- $5k–$20kMonthly recurring revenueTARGET
What is underway
In progress- MVP in buildThe radar and its read-outs, being written now
- Founder-led discoveryTalking to operators and the people who run them
- Chasing design partnersEarly companies to build with and learn from
- Pilot offer setThe operations risk audit is how we get in the door
- Outbound plannedA structured programme through Instantly.ai
Pre-seed, pre-revenue. No customers, revenue or pilots are implied — and none are claimed above.
12Team
A technical founder who can build and sell the first version.
Pre-seed and founder-led. One person, no padding.
Team · founder profileFounder-led
Founder
Trent Howes
Founder · CEO · CTO
ProductEngineeringFounder-led sales
- Writes the product. There's no one to wait on to ship a change.
- Builds with AI assistance and tunes the product from how people actually use it, not a spec doc.
- Closes the gap between spotting a problem and shipping the fix.
- Has taken software from idea to in-use before.
- Runs the sales calls and discovery, so every early user talks to the person who builds it.
Why this de-risks the plan
One person builds the product and sits in the sales calls. That keeps each early pilot cheap to win and quick to learn from: what a user struggles with on Monday can be in the build by Friday, with nobody to brief in between. The money buys shipping, not headcount.
Build loop
A problem spotted becomes a fix shipped, with no one to brief in between.
Pilot cost
The founder wins the first users, so there's no sales team to fund yet.
After the raise: first contractors and hires for engineering and design, once early pilots show the Monday brief earns its place. TARGET
Pre-seed. One founder building and selling; first hires modelled to follow the raise.
13The ask
Raising $500k–$750k pre-seed.
Enough to ship the MLP, get the first pilots running, and earn the case studies a seed round wants to see.
ASK · PRE-SEEDOpen
Target raisePLACEHOLDER
$500k–$750k
One tight round. It carries us to the point a seed investor can underwrite: a product in use, pilots paying, and a drift outcome we can put a number on.
- 18month runway
- MLPfirst milestone
- Seednext round
Use of fundsASSUMPTION
- MLP buildProduct & engineering
- IntegrationsCore connectors
- Pilots & GTMFounder-led
- Security baselineTrust & access
- First hiresContractors if needed
Indicative split — weighting will follow pilot demand.
Milestones · what this buysTARGET
- M110–30 pilotsPilots run end-to-end
- M25–15 paying customersPilots converted to paid
- M3$10k–$30k MRRRecurring revenue proven
- M42–3 publishedMeasurable operations case studies
- Every pilot is run to leave behind a before/after number on operational drift. That number is what each case study is built on.
- Hit these four and the seed conversation stops being a story and starts being evidence a follow-on round can price.
Allocation and milestone figures are modelled, not committed. No traction is implied.
In one line
Retenna helps companies see what matters before it becomes a surprise.