logo
Screenshot 2024-03-15 6.49.47 AM

Bringing the cost of data driven decisions to zero for banks and financial institutions

An expert data science AI assistant at every desk

Meet Ana

Ana is your new AI employee. She knows Masters degree level mathematics. She’s an expert Python coder and visualization builder. She believes great data analytics is about story-telling. Ana has read all your company wikis. She knows the nuances of how your company describes its important concepts. She has a To-Do list to stay on top of routine scheduled tasks. Also, you can drop by and have an ad-hoc conversation in Slack or Teams.
Ana
Clifford_sm

Meet Clifford

Clifford is Ana’s data management expert AI colleague. Clifford understands where disparate data sets are located around your business. He knows all about your existing data infrastructure investments. He has studied your enterprise catalogs and knows what's on every single management dashboard. This means he knows how to help Ana find the right data to do her work.
Watch Ana assist a bond portfolio shock scenario analysis

Why do you need us?

We can tell you something you don't already know

In most management reporting or KPI dashboarding, practitioners often already know all the information presented. This makes deriving insight very challenging.

We will alert you when your attention is needed

Decision makers rarely have the luxury to mine data for insights during their busy work days. These people need a data lieutenant to be alert across critical data changes and speak up when something important happens.

We can turbocharge your data scientists

The gap between data scientists and business users causes misunderstandings and delays. Use natural language conversations to ask questions about your data and we'll take care of the rest.

Why do customers trust us?

No hallucination impacts
All language models are capable of hallucinating - our design approach removes its impact. Ana's responses come via Python generated by our language models tuned for data analytics, not conversation. Python cannot hallucinate. If Ana receives a Python runtime error, she will stop, admit she got confused and then try again.
Explainability by design
Explanations of why a particular answer was provided by AI is often very challenging. We have made it easy. Clifford explains the concepts he understands from your questions, directs Ana to the right data. Then she generates Python before your eyes and describes what the results mean. All this is narrated in the results.
No data leakage
Your data does not get sent to a language model provider. Rather, it is queried within your own data warehouse by the SQL and Python we generate. Results are rendered in a data frame or chart. Your data is consumed in the analytics processing inside each user's session as each conversation happens. It is not sent to any third party.

How can I trust the data is good enough?

Running analytics on bad quality data can produce misleading results. Ana knows the meanings behind the data and how to use quantitative methods to detect anomalies. This means she can speak up when the data doesn't look right.

Schedule a call

Let's discuss your financial services data & analytics challenges

Frequently Asked Questions

Popular questions about TextQL
Is this intended for financial services?
Yes! Collectively, we have decades of experience working in buy-side and sell-side firms.
This sounds like science fiction. How does it work?

We are a mix of architects, engineers, dreamers, thinkers and practitioners. We are driven by bringing genuinely gamechanging technology to an industry in transformation. We are combining existing and emerging techniques to deploy advanced data modeling with language models and code generation for the purpose of solving data analytics automation. 

At the heart of TextQL is our ontology. This is how we can teach Ana and Clifford how to sound like a veteran employee at your company and not an intern!

How does Ana understand what my data means?

Public language models that have been trained on Wikipedia and lots of journal articles can display deep knowledge. But this knowledge is general, it doesn't include the specifics of how your company works such as jargon used in each department.  During implementation, we help you draw an ontology diagram in the system that explains the business nouns and important relationships. Additionally, Ana reads your Confluence wikis so she knows how you describe yourselves.

Clifford is a meta-data master. He helps Ana locate the correct data to answer questions using the precise concepts as you pose them in plain English.

What about all my management dashboards?
Ana can read those too! Power BI and Tableau dashboards may not be ideal for discovering fresh insights, but they are often a gold mine of critical metrics that describe how your company operates. They can be treated like data sources and queried in your conversations with Ana.
Is my data safe?

Yes, we take data governance very seriously. We don't store customer data in our environment, only certain meta-data to assist Clifford with retrieval. Your data remains located in your existing data warehouse platforms. As it is queried for analysis, results are encrypted in flight and at rest in each user session. A private subnet per customer disallows any other inbound traffic. Outbound IP whitelists can be configured as recommended by your technology team.

Do you have SOC II certification?
Yes, that's mandatory because we process customer data in the cloud. 
Cloud? Can we have it on-premise?
Yes. We can supply Kubernetes images if you need to host it on your site.