DecisionIQ
An agent that decides
before it writes.
Most chatbots hand every reply to a language model and hope it’s right. DecisionIQ, ShowPilot’s decision engine, works out what to show, when to ask, and when to stay quiet before the agent says a word.
Watch one engine make four different calls.
Right now
A first-time visitor asks how long setup takes
- First visit
- Came from a campaign link
- Read two product pages
- Hasn’t shared an email
- Answer, and open the install page79%
- Play the setup part of the video62%
- Ask which CRM they use44%
- Offer the setup guide31%
- Stay quiet8%
- Ask for a meetingNot ready yet
How ready this visitor is. Anything short of the line waits.Hover a move to see why.
Illustrative example. In ShowPilot you can see every decision the agent made, and why.
A chatbot replies. ShowPilot decides.
Most chatbots
Run every decision through a text-only language model. It reads the conversation, writes the most likely reply, and has no way to know whether replying was the right thing to do.
ShowPilot
Weighs who the visitor is, what they’ve read and asked, and every step the agent could take next. It picks one before a word is written.
A reply-first chatbot
With DecisionIQ
Pastes a link to the closest page it can find.
Shows the page, tour, or video that answers. Or nothing, if none does.
Says yes to a feature because the question assumed it.
Never promises what you don’t have.
Asks for the meeting on the second reply.
Asks when the buyer is ready for one.
Repeats the offer the buyer just ignored.
Backs off and gives them room.
Fills every pause with another prompt.
Stays quiet when quiet is the better move.
Guesses when it doesn’t know.
Says so, and sends the question to your team.
The five calls that win or lose a buyer.
A chatbot makes these by accident, one reply at a time. DecisionIQ makes each one on purpose.
- It only shows what fits
- A page, a tour, a video, an offer: each goes in front of a buyer only when it answers what they asked. Weak matches are left out, and the agent won’t claim a feature you don’t have.
- It asks for the meeting when the buyer is ready
- Not on the first reply, and not because a timer fired. The ask comes when the conversation has earned it.
- It backs off after an ignored nudge
- Every ask makes the next one harder to earn. When a buyer passes on an offer, the agent gives them room before trying anything like it again.
- It stays quiet when that’s best
- Someone reading your pricing page closely may not want an interruption. Saying nothing is one of the options it weighs.
- It routes what it can’t answer
- It tells the buyer your team will follow up and flags the question for you. Your answer goes out from the rep’s own mailbox.
You set the priorities. It decides when.
Tell it what’s worth most to your business: meetings, trials, or tours. DecisionIQ leans toward those outcomes in every conversation and picks the moment for each one.
- No setting can force the ask
- Put meetings at the top and it still won’t push one on a buyer who hasn’t earned it. Pushing early is how you lose them.
- Every decision is visible
- For any conversation, you can see what the agent decided and why.
- Your boundaries hold
- What you say it must never do stays off the table, however the conversation goes.
- A person can step in
- Your team can take over a live conversation and hand it back when they’re done.
No flowcharts. Your content and your rules.
DecisionIQ is only as good as what it works from. You give it your own pages and documents, and you write its goals and limits in plain English. There is no rule builder to wire up and nothing that breaks when your site changes.
It reads your site and the PDFs and snippets you add, and re-checks them on a schedule. What it says comes from what you published.
When it doesn’t know, it says so. Your team answers once, and ShowPilot emails the answer to the buyer from the rep’s own mailbox.
Questions buyers ask.
How does the agent decide what to do next?
You give it goals and rules, not lines to read. It works toward the goals you set, such as booking a meeting, in its own words for each buyer, and you can add instructions for moments that matter, such as when to open content or point at part of the page.
What signals does the AI agent read?
What the visitor says in the conversation, what they do on your site, and what your CRM knows, such as their lifecycle stage or an open deal. It notes buying signals as it talks and turns them into an intent score that ranks visitors for your team.
Does it tailor its answer to the type of buyer?
Yes. You describe your audiences in plain language, and the agent works out which one a visitor matches from the conversation and their profile, then leads with the proof you picked for that audience. There are no matching rules to build.
What do I control?
The goals, guardrails, skills and audiences the agent works with, your ideal customer profile, and which offers it may make to whom. Guardrails are hard limits it keeps even when a visitor asks.
When does it ask for a meeting?
When the visitor seems ready, not on the first message. You set when it may ask, how many times in one visit and how long it waits between asks, and it also asks at the end of a tour it presents.
Does it qualify visitors with a checklist of questions?
No. It listens for BANT, MEDDIC or MEDDPICC answers in what visitors say rather than quizzing them, and fills in a scorecard on the lead. When a routing rule needs a fact it doesn’t have, such as company size, it asks instead of guessing.
Watch it decide, live.
Book a demo and we’ll walk you through it on our demo site, and show you every call it makes.
Built by a small team that works with every customer.