AI in finance
- from the finance side, not the technology side.
AI training, advisory and implementation for controlling, accounting, accounting firms and audit firms. Led by a statutory auditor with a background in machine learning who uses AI every day - first in his own work, then with clients.
See how we workWho we work with
ConsultAI works with people who carry professional responsibility for the numbers - for whom a “20-dollar ChatGPT” solves nothing and a Big Four programme is beyond their scale.
-
Controlling
The board wants answers today, and the data sits in five files and three systems. We help controlling teams shorten the path from data to decisions: variance analysis, forecasts with ranges, management commentary in an hour instead of a day - while keeping control over what goes to the board.
-
Accounting
The largest and least looked-after group in finance: everyone talks to the CFO, yet it is accounting that answers to the tax office, the auditor, the bank and the board at once. We teach how to use AI for reconciliations, document checks, letters and reports - safely, on data that must not leak.
-
Accounting firms
Many clients, the same tasks every month and pressure on price. We show which of the firm’s processes AI can support today (Copilot, assistants, automating repetitive tasks), which it must not - and how to turn this into an advantage in conversations with clients.
-
Audit firms
The most regulated industry in finance, with the strictest confidentiality requirements. We train statutory auditors and audit teams in the practical use of AI in analytical procedures, documentation review and drafting findings - always under the auditor’s judgement, never instead of it.
Services
From awareness to implementation - step by step. Each stage makes sense on its own and prepares the next.
-
Open training and e-learning
We run open training with industry organisers - including the Education Centre of the Polish Chamber of Statutory Auditors (PIBR) and EY Academy of Business - for statutory auditors, controllers and accountants.
Read more Show less
For those who cannot make any live date, there are e‑learning courses. Current dates and sign-up: the ‘Talks and publications’ section at the bottom of the page.
-
In-house training
A programme built around your team, your tools (Copilot, ChatGPT, Claude, Gemini or local models) and your demo data.
Read more Show less
One day gives participants ready working patterns they apply the next day; two days add their own assistants and workflows. Online or on site, groups of up to 20 people, in Polish or English. The programme is built around your processes, systems and data - in days, not weeks.
-
Advisory and implementation
When training is not enough. A diagnosis of the finance process from an AI perspective (what to automate, what not to), the choice of architecture - cloud with a data processing agreement or a local model - and implementation of the first mode with a measurable result within 30 days.
Read more Show less
We start with a diagnostic sprint; further stages only if the first one has proved itself.
-
Fractional Head of AI for Finance
A fraction of a full-time role, full responsibility: someone who understands your finances and leads the AI implementation from the inside - as part of the team, not as an outside vendor.
Read more Show less
A service of limited availability: at most two clients at a time. The conversation about it starts with the diagnostic sprint.
-
Talks
Talks at industry conferences and at events companies organise for their clients and partners - from conferences to business breakfasts.
Read more Show less
45–90 minutes on what AI really does in finance, with a live demo, no hype and no robots on the slides. Dates in the “Talks and publications” section.
-
Our own tools
We build tools we use ourselves: an analyser of financial statements from XML files (Polish e‑Sprawozdania) - reading, consistency, ratios and anomalies from a single file; a weekly liquidity forecast run by an agent; assistants for finance teams.
Read more Show less
More accounting XML formats are in preparation. We make some of the tools available to clients as part of implementations.
We do not publish a price list - every programme is different. We answer every enquiry within 24 working hours: with a quote if the description of the need is complete, and if not, with three questions after which the quote takes one day.
ContactHow we work
-
80 percent finance, 20 percent technology
Read more Show less
We teach AI from the question a CFO, chief accountant or auditor asks - not from the model. Technology is the tool; the proportion is the programme.
-
Ourselves first
Read more Show less
Everything that goes into training and implementations has been used for real first - as a fractional finance manager in an industrial company, in a weekly liquidity forecasting cycle, in the documentation of our own quality system.
-
Data does not leave the office
Read more Show less
Our own on‑premise computing infrastructure (NVIDIA DGX Spark) lets us work on documents that must not go to the cloud. Three paths to choose from: free tools (for learning), cloud with a data processing agreement (for work), a local model (for sensitive data).
-
AI flags, people decide
Read more Show less
In every process the boundary is explicit: what AI prepares and what unconditionally stays with a person. We also talk about what AI will not do - and what must not be entrusted to it.
-
Quality as a system, not a promise
Read more Show less
The ConsultAI Quality Management System is certified to ISO 9001:2015 (PCC‑CERT, PCA accreditation) - implemented on our own, with the help of AI, without an implementation firm. That fact alone proves AI works where nobody expects it.
5 AI Modes for Finance
ConsultAI’s own framework that sets out what AI does in finance. Five modes are five answers to the five questions every finance director asks. The question is not “should we implement AI”, but “which mode do we implement first”.
-
DETECT
“Is there something odd in our numbers?”
Deviations from the pattern, reconciliations, anomalies in entries - flags with a materiality rating.
-
FORECAST
“What happens if…?”
Projections and scenarios always with a range - a forecast without uncertainty is false precision.
-
REVIEW
“Is this consistent, complete and material?”
Assessment of a document against criteria, a list of weighted comments - read-only mode, it never corrects on its own.
-
DRAFT
“Help me write the first version.”
Management commentary, a memo, a letter, a model - AI does 80 percent, a person the 20 percent that is the layer of responsibility.
-
ORCHESTRATE
“I do this every month - can it be automated?”
A sequence of modes with approval gates and an audit trail. It requires maturity in the first four.
About me
Grzegorz Hankus - statutory auditor with postgraduate studies in machine learning who uses AI every day
Twenty years in finance have taught me one thing: numbers have an owner, and the owner has a name. I have been a finance director in industrial companies, including during a restructuring that required rebuilding results and liquidity. I have worked in private equity. I am a statutory auditor and a member of ACCA (FCCA). To finance I added a technical education: postgraduate studies in machine learning and data engineering - which is why I know what happens under the bonnet of the tools I talk about, not just how to click them.
AI came into my work not as a topic but as a tool: first as a fractional finance manager in a portfolio company of a fund, where an experiment has to become a process, then in training - today that is over 600 trained statutory auditors and finance teams of companies, accounting firms and audit firms. In the office there is our own computing server, on which I work with documents that must not go to the cloud.
ConsultAI works in a model where one expert is responsible for the whole, and repetitive work is done by AI agents and systems - from quality documentation to the weekly liquidity forecast. For larger implementations I bring in partners to match the scope. It is a deliberate choice and at the same time proof that a serious practice can be run this way today. I safeguard quality with a system, not a promise: ISO 9001:2015, implemented with my own hands with the help of AI.
- Statutory auditor
- FCCA
- Postgraduate studies: machine learning and data engineering
- 20 years in finance (Johnson Electric, TI Automotive, TDJ)
- Fractional finance manager in an industrial company
- Author of e‑learning courses for the PIBR Education Centre
- Trainer at the PIBR Education Centre
- Trainer at EY Academy of Business
- ISO 9001:2015 QMS (PCC‑CERT, PCA accreditation, no. PW‑27807‑26)
- Training in Polish and English, online for participants anywhere in the world
Proof
- 20 years
- in finance
- over 600
- trained statutory auditors
- 70 days
- of training in 2026
- 100%
- of programmes built on real practice
References and reviews
“Mr Grzegorz combines the experience of a statutory auditor and finance manager with analytical and technological competence, thanks to which participants gain practical knowledge and concrete working patterns they can apply in practice. The PIBR Education Centre recommends Mr Grzegorz Hankus’s experience and knowledge for delivering training on AI.”
-
“It was the best training I have attended in recent years.”
- participant of a three-day on‑site training, Gdynia, July 2026 -
“I have never seen a trainer greet every participant personally, every day.”
- participant of a three-day on‑site training, Gdynia, July 2026
I work with:
- PIBR Education Centre
- EY Academy of Business
- UHY ECA
- PCC-CERT (ISO 9001) (wystąpienia)
- “Controlling i Zarządzanie” (publication)
Talks and publications
Upcoming and recent talks, publications, and the permanent places to sign up for open training or a course.
Talks
-
Conference “New technologies in financial controlling”
“The financial controller’s AI toolkit. 5 modes of work in modern controlling”
-
PIBR webinar
“3 real uses of AI in the statutory auditor’s work - step by step”
-
Conference “New competences in internal audit”
“An AI agent in internal audit - from an experiment to a tool that really works”
-
PIBR webinar
“Why ChatGPT doesn’t understand you - a practical guide to prompts for auditors”
-
“AI Business Transformation” track (Exec & VIP): “ISO 9001 without a consultancy: how AI agents built - and now run - a certified quality management system”
Recent talks
-
“AI in accounting processes. Which tools and solutions will take the load off the team?”
-
27th Annual Auditing Conference
panel “Responsibility for the future of the profession: the statutory auditor’s competences in a world of change”
Publications
-
“Business HUB”, Regional Chamber of Commerce in Katowice
“AI is already working in your company. The question is whether on your terms” - “Expert advice” column
Frequently asked questions
Is the training online or on site?
Both, but it is not the same version of the programme. Online and on site, participants work differently, so we design the exercises, pace and delivery separately for each format. On site: at the client’s premises or a chosen venue, in groups of up to 20 people.
Do participants need to know AI?
No. We assume knowledge of your own work - accounting, controlling, audit - not of technology. We adapt the programme to the level of the group: we start from the basics or go straight to a higher level if the team already has experience with AI tools. All that is needed is confident use of a computer and Excel.
Which tools are used in training?
The ones the company has: Microsoft Copilot, ChatGPT, Claude, Gemini - or local models, if the data must not leave the company. The programme has a method layer and a tool layer; the tool can be swapped.
What about data security?
We distinguish three paths: free tools (data may leak - for learning on public data only), cloud with a data processing agreement (for work), a local model on our own server (for sensitive data). In training we work on demo data; in implementations we match the path to the class of data.
How is ConsultAI different from IT firms and the Big Four?
We start from finance, not technology: it is led by a statutory auditor with 20 years of practice as a finance director and postgraduate education in machine learning, who uses AI in his own work. The scale fits mid‑sized companies and accounting firms - where a Big Four programme is beyond budget and an IT firm doesn’t understand accounting.
Where do you start implementing AI in finance?
With one mode and one process that hurts the most - with a measurable result within 30 days. Not with a 40-page strategy. A diagnostic sprint shows which process it is and what architecture to use.
Does ConsultAI run training in English and outside Poland?
Yes. In Polish and in English. Online - we run it from Poland for participants anywhere in the world; we match the time zone when planning ahead. On site - in Poland and in Europe.
How does pricing work?
We do not publish a price list because every programme is different. We answer an enquiry within 24 hours on working days: with a quote if the description of the need is complete (scope, format, date and price in one email), and if not, with three questions, after which the quote takes one day.
Let’s talk about your finances
Tell us what it is about - training, implementation, a talk - and as much as you can about the situation: who, how many people, what problem. We reply within 24 working hours: with a quote if the description is complete, or with the questions that lead to one. If you prefer to talk, leave your number - we will call back at a time that suits you.
- [email protected]
- +48 667 790 998
- Katowice