Artificial intelligence will be the top profit driver of the future. Accenture and Business Insider both said so. So did Gartner and Forrester. Everyone – from tech titans to corner pizza parlors – have been sold on the benefits of automation and machine learning…except the very industry who knows more about money than the rest of us.

Despite excitement and eagerness, banks and other financial institutions are slow to cash in on the AI bandwagon. Only a few outliers in the banking sector, like Capital One, have moved fast in launching AI-enabled products and services. We’ve heard many big banks announce ambitious plans for integrating AI and machine learning, but customers are still waiting with bated breath many months later. A recent industry survey revealed that only 32 percent of traditional financial institutions use recommendation engines, predictive analytics, voice recognition and modern machine learning technologies.

To be fair, banks are challenged by the complex regulatory and security environments they operate in. Since they manage sensitive financial and personal data for both corporations and individuals, even a data breach that affects 20 people is a matter of federal concern. An institution is still liable even if the problem was caused by a third-party failure, which is why banks have gotten more stringent about vendors hitting security standards. Many report that lack of security is the number one reason they reject a vendor, so if you’re an AI startup selling to financial services, be prepared for an intense conversation with the Chief Security Officer (CSO).

Challenges aside, the vast majority of banking executives believe AI will become the “primary” channel through which banks and their customers will interact within the next three years, according to a report by Accenture. As banks slowly warm up to AI, the winners will be the vendors who know how to navigate the complex process of selling to them.

As an AI developer, entrepreneur, or startup, here are the key steps you need to take to be enterprise-ready in the eyes of your customers:


1. Validate Business Problems Early

The number one mistake we see AI vendors making is communicating technical differentiation without focusing on specific business pain points and benefits. Does your solution solve a real problem to reduce costs and drive revenue? Schedule early meetings with customers and ask “what is your problem today?” and “what will be your problem 2-5 years from now?”. Validate the problem’s pervasiveness and negative impact on your prospect’s productivity or bottom line. Listen closely, because you might be surprised to learn that your tool could solve a different problem than the one you originally designed for.

2. Know The Laws & Regulations

Regulations in the financial industry are constantly changing, yet banks are heavily limited and motivated by them. Keep your eye on them so you can anticipate opposition that might be brought up against your solutions.

3. Be A Specialist, Not A Generalist

Don’t attempt to solve everything and be honest about what you’re not good at. High-precision industries like banking have extremely low appetite for failure and favor specialists over generalists.    

4. Overdo Your Security Layer

Getting to enterprise readiness will take longer than you think, so prepare ahead. Financial institutions perform robust “penetration tests”, or “pen tests”, where their security experts employ a range of blackhat attacks to stress test your systems. You won’t make the sale if they succeed in breaking through.

5. Choose The Right Champion

Most executives in financial institutions are risk averse and need solid proof points before adopting AI. Your challenge is to find the visionaries within those organizations who have the tech savvy, hustle, and motive to drive early AI adoption. These are often the owners of products that benefit most from machine intelligence and automation, such as call centers and consumer touch points like mobile apps. Nurture these champions with fresh intel and demos that will help them articulate your solution’s benefits to decision makers who’ll give the final word on your pitch.

6. Learn To Please Everyone

Selling to banks is a cross-department, cross-discipline affair due to an urgent need to align business drivers and technical complexities across your two organizations. Security will definitely be involved, as will Data and Information. Depending on your scope, you make also need to win over VPs, Directors and C-Suite executives in Innovation, Strategy, Marketing, and Customer Experience as well.

7. Hire Consultative Sales People

Transactional sales people fail in situations where you need to persuade a number of tech stakeholders across different departments over a period of 2-3 years. Hire sales staff that is hungry yet patient, and excellent at navigating politics and communicating with all types of personalities.

8. Join A Customer’s Startup Program

One way to shorten the long sales cycle is to join a bank’s incubator or startup program. Major financial institutions launch these innovation initiatives specifically to overcome internal inertia. Kasisto, a conversational AI platform focused on financial services, participated in Mastercard’s Start Path and Wells Fargo’s Startup Incubator and won contracts with both companies through early relationship building.

9. Ditch Your Startup Lawyer

You need a lawyer who specializes in banking and regulatory contracts. Banks will start a contract and you’ll be expected to spend a lot of time pinning down the legalease. Don’t be surprised if your first draft sets you back $10,000. If the total spend is over $3 million, you may need to add diversity agreements. If under $3 million, you may run into sustainability issues.

10. Ready An On-Premise Proof Of Concept

High security at banks means they can’t just use generic cloud providers, so you need to develop a solution that works on premise, where your software works on the bank’s servers behind their firewalls. You’ll be vying with 2 or 3 other providers for a proof of concept (PoC) and since many small startups rely on cloud providers, this is a critical point of differentiation.

11. Prepare To Run A Marathon

Closing a deal with banks can take your sales team anywhere from two to three years. That means your financial resilience and runway (and that of your investors) should match the duration of the pre-sale engagement. The procurement process alone can take 8 to 10 weeks.

12. Sprint For The Finish

Even if you win a PoC or pilot project, expect to take 2 to 3 months just to get security clearance for data access. The procurement process often takes much longer than expected, yet the bank will still have you on the hook for the originally determined launch date. Have a plan to implement your solution in a compressed time frame.


Vendors selling AI for financial services must plan, prepare, and deliver their pitch and proofs of concept flawlessly. These 12 steps will get you started, but we recommend also reading this article sharing how Kasisto successfully overcame general skepticism to win deals with major financial institutions and successfully deploy AI to production for a global bank.