Turn your best rep's playbook into workflows every rep can run
Your reps ask it to research an account, size a territory, find displacement openings, or prep for tomorrow's call. It answers in Slack or Teams, like a coworker, using HG Insights B2B data and 120+ GTM workflows.
One agent, one set of instructions your admins write. OttoBot is HG's superagent, and our own sellers used it for months before it was a product.
It works where your reps already work
Adoption is the hard part of any sales tool. So OttoBot behaves like a coworker rather than a product: your reps @-mention it in the channel they already have open, and it answers in the thread.
Slack
Mention it in the deal channel. The answer lands where the deal is already being discussed, so the whole team sees it.
Microsoft Teams
Same agent, same instructions, in the client your company already standardised on.
Prefer a full workspace? The same agent runs in the Phoenix web app, and it can be embedded in your own product for your customers.
The gap you're closing
One rep has a method that works
Your top performer researches an account a particular way. They know which signals matter, which objection lands, and what to check before a first call. It works, and it is entirely in their head.
Nobody else has it, and it leaves when they do
The method never becomes training. Enablement documents the process a quarter later, by which point it has changed. When that rep is promoted or leaves, the method goes with them.
The method becomes a workflow anyone runs
The steps your best rep takes become an agent workflow. Every rep runs it on any account, in the tool they already have open, and gets the same quality of preparation on the first call.
What every rep can run on day one
Three requests a rep actually makes. Watch one play out, including the part where the rep says the first answer isn't good enough.
A rep preps for a first call and pushes for a sharper angle · in Slack
First call with Vertex Logistics tomorrow. What should I lead with?
company_technographicVertex Logistics · full stack820 mscompany_spendVertex Logistics · 8-quarter IT spend640 mscompany_operating_signalsVertex Logistics · hiring + footprint710 ms
Pulling their stack, spend trend, and any operating signals worth opening on.
Vertex runs SAP ECC with Siemens Teamcenter for PLM and a separate Windchill install from an acquisition. IT spend is up 14% year over year, concentrated in supply-chain tooling. They're hiring six integration engineers in Rotterdam. Strongest opener: the spend increase paired with the integration hiring — they're funding a consolidation project right now.
Vertex Logistics — signal summary
- ERP
- SAP ECC (self-hosted)
- PLM
- Teamcenter + Windchill (dual)
- IT spend YoY
- +14%
- Open integration roles
- 6 (Rotterdam)
Too generic — every vendor opens on the spend number. Build it around the PLM split instead.
company_intentVertex Logistics · PLM migration topics690 mssec_full_text_search"product lifecycle" · Vertex filings1180 ms
Fair. Reframing around the dual-PLM problem specifically.
The dual PLM estate is the better wedge, and it's more defensible than the spend line. Teamcenter and Windchill both survived the 2024 acquisition, and their last annual filing names PLM consolidation as an unrealised integration synergy — so this is a commitment they've already made publicly, not a problem you're introducing. Intent data shows research on PLM migration tooling starting nine weeks ago. Open with: "Most companies carrying two PLM systems after an acquisition tell us the engineering-change process is where it hurts first." That names their situation without needing your product in the sentence.
The rep's objection was right: the revised angle cites the buyer's own filing, which a spend statistic can't do.
I'm forecasting Halden Manufacturing for this quarter. Poke holes in it before my manager does.
company_operating_signalsHalden Manufacturing · 12-month signals760 mscompany_spendHalden Manufacturing · category spend580 ms
Checking the account against the signals that usually predict slippage.
Three things weaken the forecast: 1. Their IT spend is flat over four quarters, and this category has no line of its own. 2. Headcount in the sponsoring function fell 8% since Q1 — your champion may have less scope than when you met. 3. No procurement or vendor-management signals, which usually appear four to six weeks before a contract this size closes. None of that kills the deal, but a same-quarter close needs an explanation for the missing procurement activity.
The flat spend is stale — they closed a funding round last month. Redo it with that.
company_intentHalden Manufacturing · category intent, 90d640 mssec_filing_sectionHalden · use of proceeds1040 ms
Checking what the funding actually changes.
The funding changes the spend objection but sharpens the second one. Stated use of proceeds is capacity expansion and debt refinancing — no software or digital line item. So there's fresh capital but no budget line you can point to for this purchase, which means your sponsor is competing for discretionary money rather than spending an allocated budget. That makes the missing procurement signal more concerning, not less: with allocated budget, silence often just means an efficient process. Without it, silence usually means the request hasn't been made yet. Revised read: the deal is real, the quarter is optimistic. Ask your champion whether the request is already submitted.
The rep supplied information the data didn't have. The agent took it, and reached a more pessimistic conclusion rather than an agreeable one.
Meeting the CFO at Calder Industrial next week. What does their comp plan reward?
sec_filing_sectionCalder Industrial · compensation discussion1220 ms
Reading the compensation discussion in their latest proxy.
Their annual incentive weights three measures — the split below is illustrative of the structure rather than exact figures. The short-term plan leans on margin and free cash flow; the long-term plan is relative shareholder return. So a CFO-level conversation lands better on margin durability and cash conversion than on growth.
Calder Industrial — incentive structure (illustrative)
- Adjusted operating margin
- Largest weighting
- Free cash flow conversion
- Second weighting
- Revenue growth
- Smallest weighting
- Long-term plan
- Relative TSR, 3-year
That's the CFO's own plan. Who else in the committee is measured on this, and are they aligned or fighting each other?
sec_full_text_search"segment operating" · Calder filings1090 mscompany_operating_signalsCalder Industrial · function headcount650 ms
Better question — checking how the measures cascade below the CFO.
They are not aligned, and that changes who you should be selling to. Segment leaders are measured on segment operating income, which is struck before the corporate allocations the CFO's margin measure includes. A project that raises segment cost to lower total cost helps the CFO's number and hurts theirs — so your economic buyer and your likely blocker are paid on opposing measures. The committee member whose incentives point the same way as yours is operations leadership, not the CFO: they carry cash-conversion exposure without the segment-income conflict. Suggested move: get operations into the room before the CFO meeting, so the business case arrives already reconciled.
The rep asked a question the first answer had no way to reach. Mapping incentives across the committee inverted the recommended entry point.
Territory and TAM analysis
Size a segment, find the accounts that match your best customers, and hand a ranked list to the rep who owns it.
Competitive displacement
Find who runs a competitor's product, which of those accounts show migration intent, and what to open the conversation on.
Renewal and expansion risk
Watch install-base and operating signals across your customers, and flag the accounts whose footprint is moving before renewal.
One agent to ask, a specialist per job
Your team talks to one agent. Behind it, each request is routed to a workflow built for that job, so the answers stay specific instead of becoming a general-purpose chatbot with your data attached.
One gateway
Slack, Microsoft Teams, the Phoenix web app, or embedded in your own product. Same agent, same instructions, same audit trail, whichever surface the request arrives on.
One specialist per workflow
The gateway picks the workflow that fits the request. Each specialist knows its own job and what a good answer looks like, which is why it can be challenged and revise its answer.
- Account research
- Competitive displacement
- Territory + TAM
- Renewal risk
- Exec + committee mapping
- Filing analysis
120+ GTM workflows and skills available, and the set grows as your team's methods get captured.
HG Insights B2B data, already wired in
Every specialist draws on the same data layer, so two workflows asking about one account cannot disagree about the facts. Your own systems can be added alongside it.
- Technographics
- IT spend
- Operating signals
- Intent
- SEC filings
- Firmographics
Each organisation's data, instructions, and run history stay isolated from every other organisation's.
How your motion becomes an agent
The reasonable worry about any agent is that it will sound like everyone else's. Here is the part you control.
Your instructions, not our defaults
The agent's operating instructions are files your admins edit in Phoenix. They are not a hidden system prompt: you read them, change them, and see the change take effect on the next run.
- SOUL.md
- Tone, how much it pushes back, and when it declines.
- IDENTITY.md
- Who it is to your team, and how it introduces itself on each surface.
- TOOLS.md
- Which data and tools it may reach for, and when each is appropriate.
- WORKFLOW.md
- The sequence your best rep follows, written down.
- USER.md
- Who it is working for: segments, territories, and what your team sells.
Only your admins can change it
Editing the operating instructions is restricted to organization admins. Everyone else can use the agent; the settings page turns them away.
Every run is inspectable
Sessions and sub-agent activity are surfaced in the same settings area, so when an answer looks wrong you can see the request, the workflow that handled it, and the tools it called.
SSO / OIDC
Your identity provider governs access. No separate password to manage.
Encrypted credentials
Any connection to your own systems is stored encrypted, never in a prompt.
Audit trails
Who asked what, which tools ran, and what came back. Recorded per run.
Per-org isolation
Your instructions, data, and history are scoped to your organisation.
And it keeps learning
Not the model improving on its own. Your team's methods becoming something everyone can run.
- 1
Someone improvises
A rep asks for something the workflow didn't cover, and pushes until the answer is useful.
- 2
The team copies it
The exchange is visible. Other reps start asking the same way, because it worked.
- 3
It becomes a specialist
The pattern is captured as a workflow, so it runs the same way every time instead of depending on phrasing.
- 4
The standard moves up
What your best rep did last quarter is now the floor. The next improvisation starts from there.
Why we built it this way
We learned this on ourselves. Our own sellers had been using the agent in Slack for months before it was a product, asking it to research accounts, build displacement plays, and turn the results into something they could send. The workflows on this page exist because our reps kept asking for them, not because we designed a feature list.
So the proof here is our own GTM team, not a cherry-picked customer story. Judge it the same way we did: put your instructions in and run it on accounts you already know the answers for.
Why start here instead of building it
120+ GTM workflows already shipped
The workflows exist and are running today. Building the equivalent means specifying, testing, and maintaining each one against changing data.
HG B2B data already wired in
Technographics, IT spend, operating signals, intent, and filings are already connected and consistent across workflows. Not an integration project you own.
Runs where your team already is
Slack, Microsoft Teams, the Phoenix web app, or embedded in your product. Adoption doesn't depend on anyone opening a new tool.
Put your motion in and see what it does
Talk to us about access, or work out what it would cost at your team's volume.