The trade-off behind an autopilot pipeline

11 min read

Keeping a sales pipeline current is one of those jobs everyone needs done and few people enjoy doing. Zero, a Helsinki startup, offers to take over the typing. Its customer relationship management system, or CRM, uses agents to keep deal records updated. Six customer stories describe the time saved; none times the checking that follows.

Friday needs better records

Every Friday, the sales team at Taito.ai in Helsinki was meant to go through its pipeline, the list of deals in progress. Miikka Kataja, the company’s first commercial hire, says the records had gone stale, nobody trusted them, and nobody wanted to sit through the review.

Ask about this articleHave a question? Explore the details with IC.↗

The data the team cared about lived elsewhere. Product usage sat in PostHog, a product analytics tool. Invoices were in Stripe, the payments platform, and meeting notes in Fireflies, an AI note taker. Checking a single customer meant opening PostHog and filtering by hand.

Taito’s answer was Zero, a Helsinki startup founded in 2024. Kataja’s story is one of six that Zero publishes. In Zero’s telling, the Friday review restarted because the team now trusts what it sees on screen.

That trust rests on a design choice. Zero’s agents move deals between stages and edit deal fields without asking anyone, and only outbound messages wait for a yes. The Friday meeting now depends on records the agents have changed. Someone still needs to know when to trust them.

Zero raised a $10.3 million seed round on 15 September 2026, according to a BeBeez report that names Primary Venture Partners as lead investor. A seed round is a startup’s first major outside funding.

The agents skip permission

Zero reads a team’s email, call transcripts, PostHog usage and Stripe billing, then fills in the fields of each customer record itself. At Teamspective, a leadership software company, the sales playbook lives inside the software. Before a meeting, an agent reviews what is known about the deal and flags gaps, such as who holds the budget or whether a decision deadline exists.

The product also covers prospect search, outreach sequences and customer monitoring, so customers drop other tools. Taito.ai dropped Attio (a CRM), Clay (a prospecting tool) and Lemlist (an outreach tool). Teamspective left HubSpot, Lemlist and ZoomInfo, a contact data provider. Fewer subscriptions also mean fewer connections to keep working.

Co founder Santtu Koivumäki’s guardrails post sets out which jobs agents do alone. Logging activity, enriching contacts, removing duplicates, drafting follow ups and updating deal stages and fields all happen without approval. Outbound email, LinkedIn messages and adding contacts to sequences wait for a yes. For sales that depend on judgement, the post recommends approving everything. Which mistakes a team can afford is the question seven leaders answered in our piece on when to constrain AI.

The post concedes that matching an email to a company can go wrong when domains are ambiguous, and its remedy for a mislogged call is to fix the record. Zero logs every agent action, including what changed and when. It publishes no figure for how often matches go wrong or how long checking takes. The entry has disappeared from the rep’s to-do list. The review may have taken its place.

Six customers, six accounts

Customer	Reported result	Who stands behind the figure
Taito.ai	At least 50% less time on pipeline and data maintenance	Zero’s story, from Kataja’s account
Teamspective	2x new business; every third sales hire replaced	Kantola, who labels the hire figure his estimate
Watchdog	Sales cycle cut from 60 to 30 days; 10x ARR growth in six months	Zero’s story
Pantera	8 hours saved weekly on contact management; 4x cost savings against legacy CRM tiers	Zero’s story
Sumary	5+ hours saved per active user each week; $30,000+ saved in annual HubSpot costs	Zero’s story
Emfas	400+ hours reclaimed a year; 10x more capacity for handling leads	Zero’s story

Taito’s team says it has cut the time it spends keeping pipeline data current by at least half. PostHog, Stripe and Fireflies feed each deal, and fields such as a customer’s tech stack update from the latest email or call. Handovers from sales to trial need no meetings because the context sits on the record.

Jose Kantola, Teamspective’s co founder and chief revenue officer, ran the team on HubSpot for five years and describes it as closer to data storage than a sales tool. He is also careful about where credit belongs. He calls the replaced hire his estimate and stops short of crediting Zero alone for the doubling of new business. The company sells leadership software to HR and executive teams at organisations of 500 to 5,000 people. It moved five years of history in three days, and its agents rank target companies using signals such as a chief people officer posting about a relevant topic.

Watchdog’s numbers are the most precise and come with the least context. Sebastian Ferraz, chief revenue officer at the Norwegian invoice checker, says he kept a Google Sheet beside every CRM he used because he never fully trusted the data, a habit the story places in his earlier jobs. He joined in March 2026, and the company launched in January 2026 with eight people. The story reports 10x growth in ARR, annual recurring revenue, over six months. The 60 day sales cycle describes a young product and a young team, and the story gives no deal counts.

The other three stories run shorter. Pantera, a London financial modelling platform, says the seven steps from signing a deal to onboarding the client can now be automated. Sumary, an AI finance platform founded in 2025, had its account structure running in under three weeks. Emfas moved a founder’s tracking out of Notion, a workspace tool, and into Zero within weeks. Each of the six leaves out how its figures were measured.

What else changed?

Zero’s case is straightforward. Reviewing a record an agent has filled in takes less time than writing one, and records now exist where none were kept. Taito’s Friday review is back, and Teamspective’s reps get coaching before each meeting that once depended on a manager having a spare hour. Customers report what they see, and Zero’s reading is that the savings are real even without a measurement.

The rival reading credits the other changes that arrived with the tool. Teamspective built a sales playbook while it changed systems, and writing down how a team sells can lift results with or without agents. Watchdog’s sales chief arrived in March, the same year as its CRM. Taito merged three tools and restarted a meeting at once. Small starting bases also make large growth multiples easier to achieve. Zero also chose which six customers to feature, so the stories show what is possible and say little about how common it is. The same limit applied when Sterling’s AI sales agent reported a 31.4% conversion lift without a published baseline, the starting figure a lift is measured against.

Correction time would help distinguish the two accounts: how much of the saved admin returns as review work?

The seat is cheaper

On its pricing page, checked on 10 October 2026, Zero lists the Pro plan at $60 per seat a month, shown against a struck through $77, and says the price becomes $77 after the first 12 months. A seat is one user’s licence. The page leaves open whether $60 applies to monthly or annual billing. A five person team pays $3,600 in the first year and $4,620 in the second.

Usage varies. Zero charges in credits, and each Pro seat includes 500 a month, pooled across the team. A verified email costs 1 credit, a phone number 10 and an agent AI step between 0.1 and 5. Extra credits cost $60 for 1,000 or $150 for 3,000. A five seat team’s 2,500 credits cover 2,500 emails or 250 phone numbers a month, and the agent steps that keep records current draw from the same pool. The 14 day trial comes with 2,000 credits and asks for no card.

Pairing a seat price with usage credits puts Zero in the wider debate over how AI is repricing Nordic SaaS, and a seat is a weaker measure of use once agents, not people, do the work. Sumary reports more than 5 hours saved per active user each week, about 22 hours a month. At that rate a $60 seat costs under $3 per hour saved and a $77 seat about $3.50. If the true saving were a tenth of that, around two hours a month, the seat would cost roughly $28 to $35 per hour saved, which a buyer can set against what an hour of a rep’s time costs the company. The calculation excludes review time and actual credit use, neither of which Zero publishes.

Time the checking

A buyer can measure both during the trial, and buyers should put four questions to Zero first. What share of agent stage changes and field edits do people correct, on a team like yours? Which fields feed the forecast, and can those alone require approval? How long does a correction take, and who sees that it happened? And where are email, call and LinkedIn data stored and processed? Messages and transcripts are personal data, and the pages reviewed for this piece leave open where that data goes. Prices are in dollars, so a European finance team should also confirm how VAT is applied.

The trial gives a team a chance to keep its own time sheet. Count the minutes spent maintaining records before the switch, then the minutes spent reviewing and correcting them. The Friday meeting needs dependable records, whoever enters them.

For the prospecting costs outside the CRM seat, compare Clay, Apollo and Cognism. Data enrichment and the agent that acts on those records can create separate bills.

Sources checked and limits of this piece

We read Zero’s home page, pricing page, customer index, six customer stories, a guardrails post and a product page on 10 October 2026. The funding round comes from a BeBeez report of 15 September 2026, which names Primary Venture Partners as lead investor. Industry Contents did not test the product, interview any customer or ask Zero for comment, and could not independently verify any figure in the customer stories. The pricing arithmetic uses Zero’s listed prices and Sumary’s reported saving, and it is illustrative.

FAQ

Read next

How AI workers change software economics Read the wider business implications behind this experiment.

Compare the costs and tools Use IC Lab to examine pricing models and the practical trade-offs behind AI adoption.