ToltIQ wants your deal team to stop reading in circles

A private equity deal comes with a peculiar reading assignment. The seller supplies the documents. The buyer must discover what the documents leave out, contradict or quietly qualify. A beautifully formatted presentation may sit beside a scanned contract that spoils the story. ToltIQ is built for that uncomfortable conversation between files.

Its proposition is appealingly unromantic: give investment professionals a better way to interrogate the material they already have. The customer is a private equity or credit team, a family office, a limited partner, or an advisor facing the same imbalance. There are more pages than hours. Someone still has to explain the purchase to an investment committee.

The retirement that met a PDF

Founder Ed Brandman knew the institutions before he knew this particular opportunity. He spent 2007 to 2018 as a partner and chief information officer at KKR, responsible for technology, data strategy and operations. His earlier career included trading technology at J.P. Morgan and a trading-software company of his own. He had built systems where mistakes carried consequences.

In a January 2024 founder update, Brandman credited Matthew Brandman with nudging him out of retirement to start the business. The team combined private equity and credit experience with engineering, released a purpose-built platform and reported ten firms in pilot mode. That beginning matters: the product was being tried inside an existing professional task, with actual firms, rather than waiting for someone to invent a use for it.

The company initially called itself DiligentIQ. Its June 2025 rename drew on a KKR performance review in which Brandman’s boss invoked the tölt, an additional gait of Icelandic horses. ToltIQ acquired an equestrian metaphor for efficiency. Due diligence, ordinarily short on equestrian charm, could use one.

Make the answer show its workings

ToltIQ ingests and classifies data-room material, then lets teams question it together. The platform’s practical distinction is the surrounding apparatus: shared deal context, configurable privacy and findings connected to source passages. An answer becomes something a colleague can inspect. The source link is part of the work product.

Bulk Query extracts comparable answers across documents. Workflows save a sequence of diligence steps. Vaults hold research and firm knowledge for use across deals. The point is to reduce repeated setup while keeping related work in view. Different people can examine financial, legal and commercial questions without continually rebuilding the same pile.

Blueprints, introduced in May 2026, moves the process into document creation. Firms supply an existing memo or presentation; ToltIQ converts its structure into a reusable template populated with deal information. The company says it built editing support for people and AI agents working together, alongside citations tied to actual source documents. A first draft arrives wearing the firm’s own clothes.

Spreadsheet work offers another useful example. ToltIQ’s published Data Chat test used bakery sales data to generate charts. Its advice was wonderfully specific: name the axes, the measure and the categories you want. A request for monthly sales can mean quantities or dollars. The software needs the distinction spelled out. Anyone buying AI can copy that discipline before buying anything else.

A better model is still fallible

Brandman’s March 2026 interview account includes an instructive failure. An early model, asked about the Chewy Phantom Equity Guarantee in a credit-agreement context, invented a plausible answer. He says a newer model answered correctly nine months later. It is an account of improving capability, and a reminder that fluency can arrive before knowledge.

ToltIQ’s own May 2026 research supplies a more measured picture. It tested four models, using 360 prompts from 18 financial documents and five runs per configuration. Three models showed statistically significant gains with purpose-built ingestion and retrieval. Yet synthesis-heavy tasks, including quality-of-earnings analysis, stayed below 81% across configurations. These are vendor-run benchmark results using an independent framework, not proof of a better investment.

For a buyer, the useful test follows naturally: run a completed deal through the system and inspect disputed answers against the original pages. Missing documents remain missing evidence. Conflicting clauses require interpretation. Faster retrieval cannot settle a commercial question that the record itself leaves open.

The office gets busier; the bill follows the deals

Brandman describes usage-based pricing organized around annual deal-volume tiers. That choice connects the bill to activity. ToltIQ also argues that indexing documents once and retrieving relevant material reduces wasteful repeat processing. Buyers should examine that behavior across a whole engagement, including follow-up questions and parallel workstreams.

“We don’t price on a per-user basis.”

Ed Brandman / March 2026

The financing announcement was similarly worth reading carefully: in February 2025, the company announced up to $12 million in a two-tranche Series A led by FINTOP Capital and JAM FINTOP. It identified more than 65 client organizations and named HarbourVest, Fortress, Investcorp and PPC Enterprises. H.I.G. Capital’s selection followed a two-month evaluation, according to the March 2026 announcement.

Find a place in the deal, then connect it

The market already contains financial document-analysis platforms such as Hebbia, alongside general-purpose enterprise AI. ToltIQ’s bet is that private markets expertise, document infrastructure and connected work deserve their own product. Security is part of that proposition: the company reports single-tenant isolation, SOC 2 Type II controls and no model training on client data. Its Dublin infrastructure adds a European storage option.

The integrations show where that proposition is going. Intapp connects findings to DealCloud records. PitchBook brings external comparisons into the workspace for users with an existing individual license. The August 2026 MCP release lets authorized users reach deals and vaults through Claude or ChatGPT. Each addresses a different handoff in the same process.

Those handoffs also suggest who should look elsewhere. A team needing only occasional summaries may have little reason to establish a dedicated diligence environment. A buyer expecting a PitchBook connector to supply a market-data subscription has misunderstood the offer. The case for ToltIQ becomes stronger when documents, colleagues and repeated deliverables accumulate around an active deal pipeline.

The June 2026 PwC relationship adds a larger distribution channel and a practitioner feedback loop. PwC becomes ToltIQ’s exclusive professional services advisor, with work planned around execution, client adoption and development. The lesson readers can copy is modest: choose a document-heavy task, preserve the evidence and test the complete workflow. A faster answer earns its place when someone else can check it.